CRC Press

Taylor & Francis Group

6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2020 by N.V.S Raju and BSP Publications
CRC Press is an imprint of the Taylor & Francis Group, an informa business
No claim to original U.S. Government works
Printed on acid-free paper
International Standard Book Number-13: 978-0-367-36596-7 (Hardback)
Print edition not for sale in South Asia (India, Sri Lanka, Nepal, Bangladesh, Pakistan, Myanmar or
Bhutan)
This book contains information obtained from authentic and highly regarded sources. Reasonable
efforts have been made to publish reliable data and information, but the author and publisher cannot
assume responsibility for the validity of all materials or the consequences of their use. The authors
and publishers have attempted to trace the copyright holders of all material reproduced in this
publication and apologize to copyright holders if permission to publish in this form has not been
obtained. If any copyright material has not been acknowledged please write and let us know so we
may rectify in any future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced,
transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or
hereafter invented, including photocopying, microfilming, and recording, or in any information
storage or retrieval system, without written permission from the publishers.
For permission to photocopy or use material electronically from this work, please access www.
copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC),
222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that
provides licenses and registration for a variety of users. For organizations that have been granted a
photocopy license by the CCC, a separate system of payment has been arranged.
Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are
used only for identification and explanation without intent to infringe.
Library of Congress Cataloging-in-Publication Data
A catalog record has been requested
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
 1DPR0DWKUXGHYREKDYD
1DPR3LWKUXGHYREKDYD
1DPR$FKDDU\DGHYREKDYD
1DPR$WKLGKLGHYREKDYD
$VKHHV6KLVK\DDK

Dedicated with Tributes

7R
7KH+RO\)HHWRI0\3DUHQWV

/DWH6UL190XUDOL0DQRKDU5DMX /DWH6PW$PPDD\L
3UHIDFH

6FLHQFHLVXQERXQGHGDQGQRVFLHQFHZDVERUQRQRQHGD\RUGHYHORSHGRYHUQLJKW,WLVWKHUHVXOWRI
FRQWLQXRXV GHYHORSPHQW WKURXJK FHQWXULHV 0DQ KDV EHHQ WU\LQJ WR H[SORUH WKH EHVW SRVVLEOH
PHWKRGVWRSHUIRUPKLVMREVWRUHVXOWLQWKHPRVWHIILFLHQWDQGHIIHFWLYHRXWSXWZLWKLQWKHDYDLODEOH
UHVRXUFHV2QHVXFKDWWHPSWLVPDGHWKURXJKWKLVERRN2SHUDWLRQV5HVHDUFKWRPDNHRXUVWXGHQWV
XQGHUVWDQGDQGPDVWHUWKHVXEMHFWHDVLO\
:H NQRZ WKDW PDQDJHPHQW XVHV ERWK DUW DQG VFLHQFH WR PDNH RSWLPDO GHFLVLRQV 7KHVH WZR
FRQFHSWVDUWDQGVFLHQFHDUHIXOO\HQGRUVHGWKURXJK2SHUDWLRQV5HVHDUFKDQGWKXVFDQEHFRPHD
SRZHUIXO WRRO LQ HQJLQHHU¶VPDQDJHU¶V KDQGV WR DLG KLV GHFLVLRQ SRZHU 1RW RQO\ HQJLQHHUV RU
PDQDJHUV EXW DOVR HYHU\ KXPDQ EHLQJ UHTXLUHV WKH DZDUHQHVV RI WKHVH 2SHUDWLRQV 5HVHDUFK
FRQFHSWV6RPHPLJKWKDYHEHHQHYHQXVLQJXQNQRZLQJO\EXWFDQDSSO\WKHVHPRUHFRQILGHQWO\LI
WKH\DUHIDPLOLDUZLWKWKHFRQFHSWV

2%-(&7,9(
0\ LQWHQWLRQ RI ZULWLQJ WKLV ERRN LV WZRIROG 7KH ILUVW REMHFWLYH LV WR GLVSHO WKH SKRELD WKDW
2SHUDWLRQV5HVHDUFKLVDOOPDWKHPDWLFVDQGWKDWLWUHTXLUHVDVSHFLDOSUHUHTXLVLWHRIPDWKHPDWLFDO
NQRZOHGJHZKLFKLVQRWWRWDOO\WUXH3HUKDSVWKHRULHQWDWLRQRIWHDFKLQJE\PDWKHPDWLFDOOHFWXUHUV
PLJKWKDYHJLYHQDVFRSHWRFUHDWHVXFKDIHHOLQJLQWKHVWXGHQWV7KHDSSURDFKLQWKLVERRNZHQW
RQPLQLPL]LQJPDWKHPDWLFDOGHULYDWLRQVZKLFKDUHUHTXLUHGQHLWKHUIRUSUDFWLFDODSSOLFDWLRQVQRU
IRU XQLYHUVLW\ H[DPLQDWLRQV +RZHYHU PDWKHPDWLFDO NQRZOHGJH RI VFKRRO OHYHO LV QHFHVVDU\ DQG
PRUH WKDQ VXIILFLHQW WR UHDG WKLV ERRN 0RVW RI P\ VWXGHQWV ZKR GR QRW KDYH PDWKHPDWLFDO
EDFNJURXQGDSSURDFKHGPHWRH[SODLQWKHFRQFHSWVZLWKPLQLPXPRUQRXVHRIPDWKHPDWLFV7KLV
LQIOXHQFHGPHWRDWWHPSWWRPDNHWKHLUQHHGVIHDVLEOHLQDQLQWHUHVWLQJDQGOXFLGPDQQHU
0\ VHFRQG REMHFWLYH RQ ZULWLQJ WKLV ERRN LV WR WDLORU DFFRUGLQJ WR WKH JURZLQJ QHHGV RI P\
VWXGHQWV DQG VWDII RI DOO FROOHJHV XQGHU -DZDKDUODO 1HKUX 7HFKQRORJLFDO 8QLYHUVLW\ 7KHUHIRUH D
ZLGHFRYHUDJHRIDOOWKHWRSLFVKDVEHHQRUJDQL]HGLQILYHXQLWVGLYLGHGLQWRFKDSWHUVLQWRWDO
7R PD[LPL]H WKH XWLOLW\ RI WKLV ERRN HDFK FKDSWHU LV H[KDXVWLYHO\ GHDOW ZLWK ODUJH QXPEHU RI !
 VROYHGH[DPSOHVZLWKDZLGHYDULHW\WRUHVROYHWKHFRQIXVLRQLQVWXGHQWV1RWOHVVWKDQ
SUDFWLFHSUREOHPVDORQJZLWKKLQWVDQGDQVZHUVDOPRVWDOOKDOIVROYHGLQDGGLWLRQWR! VROYHG
LOOXVWUDWLRQV PRVW RI WKHP DUH WDNHQ IURP -178 H[DP SDSHUV DUH VKRZQ ZLWK DOJRULWKPLF
DSSURDFK
$VXEREMHFWLYHRIWKLVERRNLVWRPDNHWKLVERRNDFRPSUHKHQVLYHYROXPHWRWKHIDFXOW\DOVR
+HQFHLWLVPDGHLQVXFKDZD\WKDWDQHZIDFXOW\PHPEHUWRWKLVILHOGFDQDOVROHDUQDQGWHDFK
&21675$,176
0\ VWXGHQWV RI %( % 7HFK DQG 0 7HFK RI -178 DQG RWKHU XQLYHUVLWLHV DSSURDFKHG PH WR
SURYLGH DQ H[KDXVWLYH EXW VLPSOLILHG FRQFHSWXDO DQG KDQG\ ERRN WKDW PXVW JLYH JUHDWHU WKDQ RU
HTXDOWRWKHEHVWVXFFHVVLQWKHLUXQLYHUVLW\H[DPLQDWLRQDQGDVZHOOWKDWFDQEHDSSOLHGLQWKHLUUHDO
OLIHZRUNLQJFRQGLWLRQVDIWHUWKHLUJUDGXDWLRQ%HVWHIIRUWVKDYHEHHQSXWLQWRFDWHUWKHLUQHHGVIRU

Y
YL Contents

ZKLFKPRVWTXHVWLRQVKDYHEHHQVHOHFWHGIURPWKHLUXQLYHUVLW\H[DPLQDWLRQSDSHUVDQGH[SODLQHG
FRQFHSWXDOO\ZLWKH[DPSOHVWDNHQIURPGDLO\OLIHZLWKDQHQJLQHHULQJDQGSUDFWLFDORULHQWDWLRQ

6853/86
.HHSLQJDOOWKHDYDLODEOHDQGUHTXLUHPHQWFRQVWUDLQWVLQYLHZHYHU\FKDSWHUKDVEHHQDGGHGZLWK
IORZ FKDUWV DQG ILJXUHV LQ VXUSOXV WR HDVH WKH XQGHUVWDQGLQJ 0RUHRYHU WKH VHFRQG SDJH RI HDFK
FKDSWHULVGHVLJQHGLQVXFKDZD\WRUHSUHVHQWWKHHQWLUHFKDSWHUDWDJODQFHWKDWFDQEHYHU\XVHIXO
WRVWXGHQWVIRUODVWPLQXWHSUHSDUDWLRQRIWKHLUH[DPLQDWLRQV
)XUWKHU,GRQRWFODLPWKHRULJLQDOLW\RIWKHFRQFHSWVEXWIRUWKHSUHVHQWDWLRQLQDXVHUIULHQGO\
IRUP :KLOH ZULWLQJ WKLV ERRN , ZDV EHQHILWHG LPPHQVHO\ E\ UHIHUULQJ WR PDQ\ ERRNV DQG
SXEOLFDWLRQV, H[SUHVV P\ JUDWLWXGH WRDOO VXFK DXWKRUV SXEOLVKHUV SURIHVVRUV DQG LQVWLWXWLRQV ,I
DQ\ERG\KDVEHHQPLVVHGLQDGYHUWHQWO\,VHHNKLVRUKHUSDUGRQ

&21',7,216
,QVSLWHRIWKHXWPRVWFDUHWDNHQWRPDNHWKLVERRNHTXDOWR]HURPLVWDNHVHUURUIUHHVWLOOVRPH
PLVWDNHV PLJKW KDYH JRQH RYHUVLJKW , VHHN \RXU QRQQHJDWLYH DSSURYDO IRU WKH RPLVVLRQ DQG
FRPPLVVLRQWKDWPLJKWKDYHFUHSWLQWRWKLVYROXPH$OVRDQ\VXJJHVWLRQVWRLPSURYHWKLVERRNDUH
DOZD\VXQUHVWULFWHGDQGJUDWHIXOO\DFNQRZOHGJHG

'U1965DMX
HPDLOVULMDQDWDUDM#\DKRRFRP
$FNQRZOHGJHPHQW
$WWKHRXWVHW,H[SUHVVP\GHHSVHQVHRIJUDWLWXGHIRUHQFRXUDJHPHQWDQGVXSSRUWH[WHQGHGE\«
x 'U.1DUD\DQD5DR3URI5HWGRI0(28)RUPHU0HPEHU6HFUHWDU\$,&7(1HZ
'HOKL
x 'U*9HQXJRSDO5HGG\3URIRI&6(+RQ¶EOH9LFH&KDQFHOORU-178++\GHUDEDG
x 'U195DPDQD5DR5HFWRU3URIRI&(-178++\GHUDEDG
x 'U1<DGDLDK3URIHVVRURI(((5HJLVWUDU-178+
x 'U$9LQD\D%DEX3URIRI&6()RUPHU'LUHFWRU$GPLVVLRQV-178++\GHUDEDG
)RUPHU3ULQFLSDO-178+&(+'LUHFWRU,QGXU*URXSRI,QVWLWXWLRQV
x 'U &% .ULVKQD 0XUWK\ 3URI 5HWG RI 0( -178 )RUPHU 3ULQFLSDO IRU -31(&
0DKERREQDJDU 1,(7'HVKPXNKL1DOJRQGD
x 'U 5HWG &RO . 3UDEKDNDU 5DR )RUPHU 3ULQFLSDO IRU 1,(7'HVKPXNKL 95(&
1L]DPDEDG 5DMDPDKHQGUD(QJJ&ROOHJH
,DPJUDWHIXOWRP\SUHVHQWFROOHDJXHV'U.9DVDQWKD.XPDU$30('U6XUHVK$UMXOD
$30(0UV06KDLODMD$30(-178&(-IRUWKHLUPRUDOVXSSRUWDQGPRWLYDWLRQ
, WKDQN P\ IRUPHU FROOHDJXHV 'U 1 9HQNDWHVKZDUOX $VVRF 3URI LQ 0('62(7 ,*128
0U 6RPD 6XQGDUDP .DODJD 'U % .UDQWKL .LUDQ $3&6( -178+&(- 'U '9 5DYL
6KDQNDU 3URI RI 0(  3ULQFLSDO 7.5 (QJJ &ROOHJH0U $ 6XUHVK 5DR 3URI LQ &6( 0U
1/1DUD\DQD3URIRI0(69,7DQGDOOP\SUHVHQW IRUPHUFROOHDJXHVDQGVWXGHQWV
0\VSHFLDOWKDQNVDUHGXHWRP\VWXGHQWVDQG/HFWXUHUV$GKRF0U66KLYD5DPD.ULVKQD
665. 0U 6KLYUDM &KDZDUL\D 0U 03UDGHHS .XPDU 3UHVHQWO\ $30( .,766 IRU
KHOSLQJPHLQSUHSDULQJWKHPDQXVFULSW
:KLOH DXWKRULQJ WKLV ERRN , KDG WR JR WKURXJK QXPHURXV ERRNV RQ 2SHUDWLRQV 5HVHDUFK DQG
RSWLPL]DWLRQWHFKQLTXHVDQGPHWPDQ\SURIHVVRUVSUDFWLWLRQHUV,WKDQNDOOWKRVHDXWKRUVDQGZKR
GLUHFWO\ RU LQGLUHFWO\ KHOSHG PH ZKLOH DXWKRULQJ WKLV ERRN ,I IHHO VRUU\ LI , PLVVHG DQ\
LQDGYHUWHQWO\
, IXUWKHU DFNQRZOHGJH WKH FRPSXWHUL]DWLRQ WHDP RI WKLV ERRN DQG DFWLYH IROORZ XS DQG
SURFHVVLQJE\WKHSXEOLVKHUV6UL1LNKLO6KDKDQG6UL$QLO6KDK7KHZRUNRI0U09DVXGHYD
5DR 0U 1DUHVK 0V 6DQGK\D DQG 0U +DUL 3UDVDG DUH SUDLVHZRUWK\ GXULQJ WKH SXEOLVKLQJ
SURFHVV
/DVWEXWQRWWKHOHDVW,WKDQNP\ZLIH0UV3UDVDQQDDQGFKLOGUHQ.XP6ULMD1DWDUDM.XP
+LPDMD 3DGPD 3UL\D DQG .XP 6DQMD .DUXQDPD\L IRU WKHLU FRRSHUDWLRQ  SDWLHQFH DQG P\
3DUHQWV6PW 6UL/DWH190XUDOL0DQRKDU5DMXDQG%URWKHUV'U9HQNDWD15DMXDQG
0U196ULQLYDVD5DMXDQGVLVWHUV6PW+HPDODWKD6PW3UHPDODWKDDQG6PW6XPDODWKD
IRU WKHLU HQFRXUDJHPHQW )LQDOO\ , WKDQN HQWLUH .DVX DQG &KHQQDPDGKDYXQL IDPLOLHV IRU
VKRZHULQJWKHLUJRRGZLVKHVRQPH

'U1965DMX

YLL
&RQWHQWV
Preface ............................................................................................................................................................ Y
Acknowledgement .......................................................................................................................................... YLL

&+$37(5
2SHUDWLRQV5HVHDUFK±$Q2YHUYLHZ
 ,QWURGXFWLRQ
 2ULJLQDQG'HYHORSPHQWRI250RGHOV
 2SHUDWLRQV5HVHDUFK6RPH'HILQLWLRQV
 3KDVHVRU6WHSVRI2SHUDWLRQ5HVHDUFK0HWKRG0RGHOLQJ
 2SHUDWLRQ5HVHDUFK0RGHOV&ODVVLILFDWLRQ
 &ODVVLILFDWLRQ%DVHGRQ6WUXFWXUH
 &ODVVLILFDWLRQ%DVHGRQ3XUSRVHDQG1DWXUH
 %DVHGRQ&HUWDLQW\
 %DVHGRQ7LPH5HIHUHQFH
 %DVHGRQ0HWKRGRI6ROXWLRQ
 $SSOLFDWLRQ0RGHOVRI2SHUDWLRQ5HVHDUFK
 6LJQLILFDQFHDQG$SSOLFDWLRQVRI25LQ,QGXVWULDO3UREOHPV
 &KDUDFWHULVWLF)HDWXUHVRI2SHUDWLRQV5HVHDUFK0RGHO
RU0HULWVRI2SHUDWLRQV5HVHDUFK0RGHOV
 'LVWLQFW'LUHFWLRQLQ'HFLVLRQ0DNLQJ
 6FLHQWLILF$SSURDFK
 2EMHFWLYH2ULHQWDWLRQ
 ,QWHU'LVFLSOLQDU\7HDP$SSURDFK
 &RPSDWLEOHWR'LJLWDO&RPSXWHU
 &RQVLGHUDWLRQRI$OO)DFWRUV
 /LPLWDWLRQV'HPHULWVRI2SHUDWLRQV5HVHDUFK
 ,PSRVVLELOLW\DQG,QIHDVLELOLW\
 )RU,QWDQJLEOH6LWXDWLRQV
 'LVWDQFH%HWZHHQ0DQDJHUV'HFLVLRQ0DNHUDQG
2SHUDWLRQ5HVHDUFKHU
5HYLHZ4XHVWLRQV
2EMHFWLYH7\SH4XHVWLRQV
)LOOLQWKH%ODQNV
$QVZHUV

L[
[ Contents

&+$37(5
)RUPXODWLRQRI/LQHDU3URJUDPPLQJ3UREOHPV
 ,QWURGXFWLRQ
 )RUPXODWLRQ
 0DMRU$VVXPSWLRQRI//3
3UDFWLFH3UREOHPV
5HYLHZ4XHVWLRQV
2EMHFWLYH7\SH4XHVWLRQV
)LOOLQWKH%ODQNV
$QVZHUV

&+$37(5
*UDSKLFDO6ROXWLRQRI/33
 ,QWURGXFWLRQ
 0HULWVDQG'HPHULWVRI*UDSKLFDO6ROXWLRQV
 *UDSKLFDO6ROXWLRQ3URFHGXUH
 5HGXQGDQW&RQVWUDLQW
 'LIIHUHQW7\SHVRI6ROXWLRQV
 /RFDWLQJ)HDVLEOH5HJLRQ
 7\SHVRI)HDVLEOH5HJLRQV
 7\SHVRI2SWLPDO6ROXWLRQV
6ROYHG3UREOHPV
3UDFWLFH3UREOHPV
$GGLWLRQDO3UREOHPV
5HYLHZ4XHVWLRQV
2EMHFWLYH7\SH4XHVWLRQV
)LOOLQWKH%ODQNV
$QVZHUV

&+$37(5
6LPSOH[0HWKRG
 ,QWURGXFWLRQ
 6LPSOH[$OJRULWKP
 6RPH6SHFLDO&DVHV
 8QERXQGHG6ROXWLRQ
 0XOWLSOH2SWLPDO6ROXWLRQV
 8QLTXH6ROXWLRQ
 ,QIHDVLEOH6ROXWLRQ
3UDFWLFH3UREOHPV
$GGLWLRQDO3UREOHPV
 Contents [L

 6LPSOH[0HWKRG8VLQJ6XUSOXV $UWLILFLDO9DULDEOHV

 %LJ00HWKRG
 7ZR3KDVH0HWKRG
 'HJHQHUDF\LQ6LPSOH[0HWKRG
 7LHIRU(QWHULQJ9DULDEOHLH.H\&ROXPQ
 8QUHVWULFWHG9DULDEOHVLQ//3
 (TXDOLW\&RQVWUDLQWVLQ//3RU0L[HG&RQVWUDLQWV
 &RQFHSWRI'XDOLW\LQ6LPSOH[
 &RQYHUVLRQRI3ULPDOWR'XDORU'XDOWR3ULPDO
 &RPSDULVRQEHWZHHQ3ULPDO 'XDO
 $GYDQWDJHV $SSOLFDWLRQVRI'XDOLW\
 6KDGRZ3ULFH
3UDFWLFH3UREOHPV
$GGLWLRQDO3UREOHPV
3UDFWLFH3UREOHPVLQ'XDOLW\
5HYLHZ4XHVWLRQV
5HYLHZ4XHVWLRQVLQ'XDOLW\
2EMHFWLYH7\SH4XHVWLRQV
)LOOLQWKH%ODQNV
$QVZHUV

&+$37(5
7UDQVSRUWDWLRQ3UREOHP
 ,QWURGXFWLRQ
 )RUPXODWLRQ
 7UDQVSRUWDWLRQ3UREOHP)RUPXODWHGDVDQ/33
 7UDQVSRUWDWLRQ$OJRULWKP6ROXWLRQ0HWKRG
 0D[LPLVDWLRQ&DVHLQ7UDQVSRUWDWLRQ3UREOHPV
 8QEDODQFHG7UDQVSRUWDWLRQ3UREOHP
 'HJHQHUDF\LQ7UDQVSRUWDWLRQ3UREOHP
 8QEDODQFHG7UDQVSRUWDWLRQ3UREOHP
 1RUWK:HVW&RUQHU0HWKRG
 9RJHO¶V$SSUR[LPDWLRQ0HWKRG9$0RU3HQDOW\0HWKRG
 /HDVW&RVW(QWU\0HWKRG0DWUL[0LQLPD0HWKRG
 1RUWK:HVW&RUQHU0HWKRG
 6WHSSLQJ6WRQH0HWKRG
 02',0HWKRG
[LL Contents

 9RJHO¶V$SSUR[LPDWLRQ0HWKRG9$0RU3HQDOW\0HWKRG

 'HJHQHUDF\LQWKH,QLWLDO6ROXWLRQ
 'HJHQHUDF\LQWKH6XEVHTXHQW,WHUDWLRQV
 /HDVWFRVW(QWU\0HWKRG0DWUL[0LQLPD0HWKRG
 0XOWLSOH2SWLPDO6ROXWLRQRU$OWHUQDWH2SWLPDO6ROXWLRQ
 5HVWULFWHG73
 &RQGLWLRQDO73
3UDFWLFH3UREOHPV
 7UDQVVKLSPHQW3UREOHP
5HYLHZ4XHVWLRQV
2EMHFWLYH7\SH4XHVWLRQV
)LOOLQWKH%ODQNV
$QVZHUV

&+$37(5
$VVLJQPHQW3UREOHP
 ,QWURGXFWLRQ
 7KH$VVLJQPHQW$OJRULWKP
 9DULDWLRQVLQ$VVLJQPHQW3UREOHP
 0D[LPLVDWLRQ&DVHLQ$3
 8QEDODQFHG$VVLJQPHQW3UREOHP
 0XOWLSOH2SWLPDO6ROXWLRQ
 5HVWULFWHGRU3URKLELWHG$VVLJQPHQW
 &RQGLWLRQDO$VVLJQPHQW
 &UHZ$VVLJQPHQW3UREOHP
 7UDYHOOLQJ6DOHVPDQ3UREOHPRU5RXWLQJ3UREOHP
 $7UDQVSRUWDWLRQ3UREOHPDV$VVLJQPHQW3UREOHP
3UDFWLFH3UREOHPV
$GGLWLRQDO3UREOHPV
5HYLHZ4XHVWLRQV
2EMHFWLYH7\SH4XHVWLRQV
)LOOLQWKH%ODQNV
$QVZHUV

&+$37(5
,QYHQWRU\0DQDJHPHQW
 ,QWURGXFWLRQ
 &ODVVLILFDWLRQRI,QYHQWRU\
 7KH2SHUDWLQJ'RFWULQH
 7KH,QYHQWRU\&RVWV
 Contents [LLL

 7HUPLQRORJ\XVHGLQ,QYHQWRU\
 ,QYHQWRU\6LJQLILFDQFH$GYDQWDJHVRI,QYHQWRU\
 2EMHFWLYHVRI,QYHQWRU\
 )XQFWLRQVRI,QYHQWRU\
 )DFWRUV$IIHFWLQJ,QYHQWRU\
 &RQFHSWRI(FRQRPLF2UGHU4XDQWLW\(24
 ,QYHQWRU\0RGHOV
 0RGHO:LOVRQ+DUULV6LPSOH(24
 0RGHO&RQWLQXRXVDQG1RQXQLIRUP'HPDQG&RQVXPSWLRQ
3UDFWLFH3UREOHPV
 0RGHO:LWK&RQWLQXRXV3URGXFWLRQ6XSSO\5DWH
3UDFWLFH3UREOHPV
 0RGHO'HWHUPLQLVWLF0RGHOZLWKVKRUWDJHV
3UDFWLFH3UREOHPV
 0RGHO:LWK)LQLWH5HSOHQLVKPHQWDQG6KRUWDJHV$OORZHG
3UDFWLFH3UREOHPV
 0RGHO:LWK3ULFH%UHDNVRU4XDQWLW\'LVFRXQWV
3UDFWLFH3UREOHPV
 6WRFKDVWLFRU3UREDELOLVWLF0RGHOVRI,QYHQWRU\&RQWURO
 0RGHO6LQJOH3HULRG3UREDELOLVWLF0RGHOZLWK,QVWDQWDQHRXV
'HPDQGZLWKRXW6HWXS&RVW
3UDFWLFH3UREOHPV
 ,QYHQWRU\&RQWURO$SSURDFKHV
 6WRUHV0DQDJHPHQW 6HOHFWLYH,QYHQWRU\&RQWURO
5HYLHZ4XHVWLRQV
2EMHFWLYH7\SH4XHVWLRQV
)LOOLQWKH%ODQNV
$QVZHUV

&+$37(5
4XHXH0RGHOV:DLWLQJ/LQHV
 ,QWURGXFWLRQ
 7HUPLQRORJ\
 :KHQ'RHV$TXHXH5HVXOW"
 &RVWV$VVRFLDWHGZLWK:DLWLQJ/LQHV4XHXLQJ
 4XHXH6\VWHP6DOLHQW)HDWXUHV
 4XHXLQJ0RGHOV.HQGDOO/HH1RWDWLRQV
 3DUDPHWHUVRI4XHXH6\VWHP
 $[LRPVRI3RLVVRQ3URFHVV
[LY Contents

 2SHUDWLQJ&KDUDFWHULVWLFVRI4XHXH6\VWHP
 0RGHO00’),)2
3UDFWLFH3UREOHPV
 0RGHO00’6,52
 02'(/001)&)6
3UDFWLFH3UREOHPV
 0RGHO006’),)2
3UDFWLFH3UREOHPV
5HYLHZ4XHVWLRQV
2EMHFWLYH7\SH4XHVWLRQV
)LOOLQWKH%ODQNV
$QVZHUV

&+$37(5
6HTXHQFLQJ
 ,QWURGXFWLRQ
 6HTXHQFLQJRIQ-REV;0DFKLQH
 6HTXHQFLQJRIQ-REV;0DFKLQHV-RKQVRQ¶V5XOH
 $VVXPSWLRQVIRUQ-REV;0DFKLQHV6HTXHQVLQJ
 7KH5XOH
3UDFWLFH3UREOHPV
 6HTXHQFLQJQ-REV;0DFKLQHV-DFNVRQ¶V5XOH
 6HTXHQFLQJQ-REV;P0DFKLQHV
3UDFWLFH3UREOHPV
 6HTXHQFLQJRI-REV;Q0DFKLQHV
3UDFWLFH3UREOHPV
5HYLHZ4XHVWLRQV
2EMHFWLYH7\SH4XHVWLRQV
)LOOLQWKH%ODQNV
$QVZHUV

&KDSWHU
5HSODFHPHQW$QDO\VLV
 ,QWURGXFWLRQ
 )DLOXUH3DWWHUQV
 'HILQLWLRQ
 %HKDYLRXURI0DFKLQHV
 )DLOXUHVEDVHGRQWKH9ROXPHRI)DLOXUH
 7\SHVRI)DLOXUHVEDVHGRQWKH0RGHRI)DLOXUH
 Contents [Y

 )DLOXUH&RVWV
 7\SHVRI5HSODFHPHQW3UREOHPV
 7\SH5HSODFHPHQW3ROLF\ZKHQ0RQH\9DOXHGRHVQRW
&KDQJHZLWK7LPH
3UDFWLFH3UREOHPV
 5HSODFHPHQW3ROLF\IRULWHPV:KHQ0RQH\9DOXH&KDQJHVZLWK7LPH
3UDFWLFH3UREOHPV
 *URXS5HSODFHPHQW
3UDFWLFH3UREOHPV
5HYLHZ4XHVWLRQV
2EMHFWLYH7\SH4XHVWLRQV
)LOOLQWKH%ODQNV
$QVZHUV

&+$37(5
*DPH7KHRU\
 ,QWURGXFWLRQ

 7KHRU\RI*DPHV7KH7HUPLQRORJ\XVHG

 3URSHUWLHVRID*DPH
 'HILQLWLRQVRI7HUPVXVHG
 7\SHVRI*DPHV
 :LWK5HIHUHQFHWR1XPEHURI3OD\HUV
 :LWK5HIHUHQFHWR1DWXUHRI2XW&RPHV
 :LWK5HIHUHQFHWR&HUWDLQW\
 :LWK5HIHUHQFHWR9DOXHRIWKH*DPH
 7ZR3HUVRQ]HUR6XP*DPHVRU5HFWDQJXODU*DPHV
 $VVXPSWLRQVRIWKH*DPH
 6ROYLQJ7ZR3HUVRQ]HUR6XP*DPHV
 'RPLQDQFH3ULQFLSOH
 *DPHVZLWKRXW6DGGOH3RLQWV0L[HG6WUDWHJ\
 *UDSKLFDO6ROXWLRQV
 /LQHDU3URJUDPPLQJ0HWKRG
 'LVWLQFWLRQEHWZHHQ'HWHUPLQLVWLFDQG3UREDELOLVWLF*DPHV
 /LPLWDWLRQVRI*DPH7KHRU\
 )ORZ&KDUWRI*DPH7KHRU\
3UDFWLFH3UREOHPV
5HYLHZ4XHVWLRQV
2EMHFWLYH7\SH4XHVWLRQV
)LOOLQWKH%ODQNV
$QVZHUV
(xvi) Contents

CHAPTER 12
Dynamic Programming
12.0 Introduction ...................................................................................................................619
12.1 The Approach & Algorithm ..........................................................................................619
12.2 Salient Features & their Definitions of DPP .................................................................620
12.3 Characteristics of DPP ..................................................................................................621
Practice Problems ..........................................................................................................636
Review Questions..........................................................................................................640
Objective Type Questions .............................................................................................640
Fill in the Blanks ...........................................................................................................642
Answers .........................................................................................................................642
 IS)

Collection of mathematical techniques

and tools to solve practical decision problems
Science that aids decision making
Art of giving bad decisions otherwise
which would have been the worst
Scientific and systematic approach of problem
solving for executive management

Impossibility and infeasibility in certain

- Distinct directions in decision making
real life or practical problems - Scientific approach
Not applicable to intangible variables - Objective orientation
and attributes such as human factors - Inter-disciplinary team approach
Gap between decision maker (manager) - Compatible to digital computer
and operations researchers - Consideration of all factors

Identification of me problem
A total system orientation Operations Collection of relevant data > 1
Scientific approach
Research Formulation of data into mathematical form o
Realistic nature
Application of suitable OR model ir-1
Models quantifiable 0)
Analysis and calculation
Systematic method
Decision making <D

8
Allocation models (simpiex/TP/AP) - Physical (Iconic / Analogue) Models/
Inventory (Deterministic /Probabilistic) symbolic (Verbal / Mathematical)
Game theory 0
- Descriptive / Predictive /Normative
Network models (CPU I PERT) o
- Deterministic / Probabilistic
Sequencing models 30)
Queue models - Static / Dynamic
Q.
Replacement - Analytical / Iterative/Simulative
• Dynamic programming •o
3
Operations Research - An Overview 3

1.0 Introduction
"Science that aids decision making is Operations Research which is tested and
implemented through appropriate modelling."

We take several decisions in our daily life. Most of these decisions are taken by
common sense. But the decisions taken by mere common sense sometimes may
mislead or confuse us. Also, such decisions may neither provide any evidential support
nor stand on any scientific base. Therefore, it has become necessary for managers and
engineers to believe in the science that provides the evidential support and scientific
base.

Operations Research (OR) is one such science that provides better solutions to
the managers, engineers and any practitioners with better solutions. This science came
into existence during World War II. Though it was first employed for military
operations, its applications are extended to any field on the earth in some form or
other.

Thus OR is considered as the science that deals with decision making and this
book deals with formulating, analysing, testing and application of various OR models.

1.1 Origin and Development of OR Models

The term, "Operations Research" was first coined by Me Closky and Trefthen in
1940 in a small town, Bowdsey of United Kingdom.

The name operations research was given to this subject because it has started
with the research of (military) operations. During world war - II, the military commands
of UK and USA engaged several teams of scientists to discover tactical and strategic
military operations. Their mission was to formulate specific proposals and to arrive
to the decisions that can optimally utilize the scarce resources to acquire maximum
possible level of effective results. In simple words, it was to uncover the methods that
can yield greatest results with little efforts. Thus it has gained popularity and was called
"an art of winning the war without actually fighting if.

Following the end of the war, the success and encouraging results of British
teams have attracted industrial managers to apply these methods to solve their
complex problems. The first method in this direction was simplex method (LPP)
developed in 1947 by G.B. Dantzig, USA. Since then several scientists have been
developing this science in the interest of making operations to yield high profits or
least costs.

Now, this science has become universally applicable to any area such as
transportation, hospital management, agriculture, libraries, city planning, financial
institutions, construction management and so forth. In India, many industries have
been realising the advantages by implementing the OR models.
 Operations Research : Theory and Practice

A few to quote in this regard are Delhi Cloth Mills, Indian Airlines, Indian
Railways, Hindustan Liver Ltd., (HLL), Tata Iron & Steel Co., (TISCO), Fertilizer
Corporation of India (FCI), Life Insurance Corporation (LIC) of India etc.

1.2 Operations Research : Some Definitions

Because of the wide scope of applications of operations research, giving a
precise definition is difficult. However, a few definitions of OR are as under.
• Operational research is the application of the methods of science to complex problems, in
the direction and management of large systems of men, machines, materials and money
in industry, business, government and defense. The distinctive approach is to develop a
scientific model of the system incorporating measurements of factors such as change and
risk, with which to predict and compare the outcome ofalternative decisions, strategies or
controls. The purpose is to help management in determining its policy and actions
scientifically. Operations Research Society, UK
• Operations Research is concerned with scientifically deciding how to best design and
operate man - machine systems usually requiring the allocation of scarce resources.
Operations Research Society, America

Apart from being lengthy, the definition given by Operational Research Society
of UK, has been criticised because it emphasizes complex problems and large systems,
leaving the reader with the impression that it is a highly technical approach suitable
only to large organisation. The definitions of OR society of America contains an
important reference to the allocation of scarce resources. The key words used in the
above definitions are scientific approach, scarce resources, system and model. The British
definition contains reference to optimisation, while the American definition quietly
slips in the word., best. A few other definition which are commonly used and widely
acceptable are as follows :
• Operations Research is the systematic application ofquantitative methods, techniques and
tools to the analysis ofproblems involving the operation of systems.
Daellenbach and George, 197 8

• Operations Research is essentially a collection ofmathematical techniques and tools which

in conjunction with a systems approach, are applied to solve practical decision problevis
of an economic or engineering nature. Daellenbach and George, 1978
These two definitions project another view of OR as the collection of models
and methods which have grown up largely independent of one another.
• Operations Research utilizes the planned approach (updated scientific method) and an
interdisciplinary providing a quantitative basis for decision making and uncovering new
problevis of quantitative analysis. Thierauf and Klekamp, 1975
• This new decision making field has been characterised by the use of scientific knowledge,
through inter disciplinary team effort for the purpose of determining the best utilisation
of limited resources. H.ATaha, 1976
These two definitions refer to the inter disciplinary nature of OR. However,
one of the best definitions, given by Churchman, Ackoff and Arnoff, is as follows :
Operations Research - An Overview 5

• Operations Research, in the most general sense, can be characterised as the application
of scientific methods, techniques and tools, to problems involving the operations ofa system
so as to provide those in control of the operations with optimum solutions to the problems.
Churchman, Ackoff and Arnoff, 1957

• Operation Research has been described as a method, an approach, a set of techniques, a

team activity, a combination ofviany disciplines, an extension of particular disciplines
(mathematics, engineering and economics), a new discipline, a vocation, even a religion.
It is perhaps some of all these things. S L Cook, 1977
• OR is the art of giving bad, answers to problems to which otherwise worse answers are
given. TLSaaty, 1958
• Operations Research may be described as a scientific approach to decision making that
involves the operations of organisational system. F S Hiller and G J Lieberman, 1980
• Operations Research is a scientific method of providing executive departments with a
quantitative basis for decisions under their control. P M Morse and G E Kimball
• Operations Research is applied decision theory. It uses any scientific, mathematical, or
logical means to attempt to cope with the problems that confront the executive, when he
tries to achieve a through - going rationality in dealing with his decision problems.
D W Miller and M K Stan

• Operations Research is a scientific approach to problems solving for executive

management. H M Wagner
As the discipline of operations research grew, many other names such as
Optimization Techniques, Operation Analysis, Systems Analysis, Decision Analysis,
Management Science were given to it. However, each of these emphases the
quantitative approach to the analysis and solution of management problems.

1.3 Phases (or Steps) of Operation Research Method (Modelling)

About four to five decades ago it would have been difficult to get a single
operations researcher to describe a procedure for conducting OR project. But today,
every engineer and manager are presenting the procedure in different ways as suits
to their situations. However, the procedure for an OR study generally involves the
following major phases :
Phase - 1 : Formulating the problem
Phase - I I : Fitting a Suitable 'OR' model
Phase - III : Analysis and Deriving the solutions from the model
Phase - IV : Testing the Model and its solution
Phase - V : Sensitivity analysis and Controlling the solution
Phase - VI: Decision making and Implementing the solution
Phase -1 : Formulating the Problem :
To find the solution of a problem, the problem must be formulated in the form
of an appropriate model. This requires the following information.
1. Who has to take the decision?
2. What are the objectives?
3. What are the ranges of controllable variables?
 Operations Research : Theory and Practice

4. What are the uncontrolled variables that may affect the possible solutions?
5. What are the restrictions (or) constraints oh the variables?
6. What are other conditions or nature of variables?
The formulation should be considerably careful while executing this phase
because a wrong formulation cannot give a right decision (solution), and even may be
disastrous in some cases.

Phase - II : Fitting a Suitable (OR' Model:

The next phase of the investigation is concerned with the searching or suitable
OR model in an appropriate form which is convenient for analysis. It requires the
identification of both static and dynamic structural elements. The OR model consists
of the following three important basic factors.
1. Decision variables and OR parameters:
2. Constraints (or) restrictions;
3. Objective function.
Various models are discussed in the next sections (1.4 & 1.5)
Phase - III : Analysis and Deriving the Solutions from the Model:
The third phase involves in the computation of those values of decisions
variables that maximise or minimise the objective function. Such solution is called an
optimal solution which is always in the best interest of the.problem under
consideration.

Phase - IV : Testing the Model and its Solution :

After computing and deriving the solution from the model, it is once again
tested as a whole for the errors if any. A model may be said to be valid if it can provide
a reliable prediction of the system's performance. A good practitioner of Operations
Research believes that his model be applicable for a longer time and thus updates the
model time to time by considering the past, present and future specifications of the
problem.

Phase - V : Sensitivity Analysis and Controlling the Solution :

The fifth phase establishes controls over the solution with desired degree of
satisfaction. The model requires immediate modification as soon as the controlled
variables (one or more) change significantly, otherwise the model goes out of control.
This action is often referred to as sensitivity analysis. As the conditions are constantly
changing in the world, the model and the solution may not remain valid for a long
time. Therefore it leaves a lot of scope for further improvement along with controlling.

Phase - VI : Decision Making and Implementing the Solution :

Finally, the tested results of the model are implemented to work. This phase is
primarily executed with the co-operation of operations Research experts and those
who are responsible for managing and operating the systems. However, this job is
more concerned with the production dept of executors of the model. Moreover, it
essentially requires to take into confidence, the willingness and belief of the people
who implement it or involve in it.
Operations Research - An Overview 7

1.4 Operation Research Models: Classification

OR model is a representation of a real life situation. OR models can be studied
by classifying in many ways. These are discussed here below :

1.41 Classification Based on Structure

1. Physical Models: These models give a physical appearance of the real object
either in reduced form or scaled up. These are further divided into two
categories.
(a) Iconic Models : These are representations in either idealised form or a
scaled version, (i.e., enlarging or reducing in size) of real objects, e.g., Blue
prints, Globe, Photographs, Drawings, Templates etc.

(b) Analogue Models : These models represent a system by a set of properties

different from that of the original system, and physically do not resemble.
After attaining solution, it is re-interpreted in terms of original system, e.g.,
frequency curves, flow charts, organisation charts etc.
Classification of OR Models

Based on Based on Based on Based on Based on

structure purpuse/nature degree of certainity time reference method of solution

Descriptive Predictive Normative Static Dynamic

r
Physical Symbolic Deterministic Probabilistic
I J \
Analytical Iterative Simulative
j i or stochastic

Iconic Analog Verbal

1
Mathematical

FIGURE 1.1 : CLASSIFICATION OF OR MODELS

2. Symbolic Models : These models use symbols in the form of letters,
mathematical operators or any other symbols to represent the properties of
the system. These are often described in two types.
(a) Verbal Models : These models are used to describe a situation in written or
spoken language in the form of letters, words and sentences.
(b) Mathematical Models : These models use mathematical symbols, letters,
numbers and mathematical operators to represent relationship among
various variables of the system to explain its behaviour and properties.

1.4.2 Classification Based on Purpose and Nature

1. Descriptive Models : The reports of surveys, questionnaire results,
inferences of the observations etc., are used in such models to describe the
situation. These also include the models such as plant layout diagram, block
diagram of an algorithm etc.
 Operations Research : Theory and Practice

2. Predictive Models : These models are results of the quiery such as "what will
follow if this occurs or does not occur?" e.g. Preventive maintenance
schedules.
3. Normative Model (or Optimisation Models) : These models are designed to
provide 'optimal' solution to the problems subject to certain limitations on
the use of resources or meeting the requirements or at the conditions that
normally exist, e.g. Linear Programming Problem.

L4.3 Based on Certainty

1. Deterministic Model : If all the parameters of decision variables, constants
and their functional relationship are known (or assumed to be known) with
certainty, then the model is said to be deterministic, e.g. Certain inventory
models, games with saddle points.
2. Probabilistic (or Stochastic) Model : This is the model in which at least one
parameter or decision variables is a random variable, e.g. Probabilistic
inventory models, games without saddle points, queuing models.

1.4.4 Based on Time Reference

1. Static Model : These models present a system at a specified time, which do
not account for changes over certain period of time. e.g. Replacement of
machines when money value is not changing with time.
2. Dynamic Model: Time is considered as one of the variables and the impact
of changes generated by time is accounted while selecting the optimal
course of action, e.g. Machine replacement model when money value is
changing with time.

L4.5 Based on Method of Solution

1. Analytical Model : These have a specific mathematical structure and hence
can be solved by analytical or mathematical techniques, e.g. Any
optimisation model, such as inventory models, waiting lines etc.
2. Iterative Models : In these models, the solution is obtained from the
conclusion of previous step. e.g. Simplex method for LPP, dynamic
programming.
3. Simulation Models : Though these models have a mathematical structure,
they can not be solved by applying mathematical techniques. Instead, a
simulation model is essentially a computer assistant experimentation on a
mathematical structure of a real life problem under certain assumptions over
a period of time. e.g. Monte-Carlo simulation, use of random numbers,
forecasting models etc.

1.5 Application Models of Operations Research

Basic Operation Research Application Models are discussed here below
1. Allocation Models : Allocation models are used to allocate resources to
activities in such a way that sum measure of effectiveness (objective function)
is optimized, e.g. Linear Programming Problem (Simplex, Transportation,
Assignment etc.,) and Non-linear programming.
Operations Research - An Overview 9

2. Inventory Models : Inventory models are used to determine how much to

order and when to place an order so as to optimize the inventory costs such
as order cost, carrying cost, shortage cost etc.,
3. Competitive (Game Theory) Models : These models are used to characterize
the behaviour of two or more opponents (called players) who compete for
the achievement of conflicting goals. These models are classified according
to number of competitors, sum of loss and gain and type of strategy.
4. Network Models : These models are applied to management (planning and
scheduling) of large scale projects, PERT/CPM help in identifying trouble
spots in a project by critical path and to determine time-cost trade off,
resource allocation and updating of activity times.
5. Sequencing Models : This arises whenever there is a problem in finding a
sequence or order in a number of tasks performed by a number of service
facilities. Here, the total time to process all the jobs on all the machines is
optimized.
6. Waiting Line (Queuing) Models : These are used to establish a trade off
between cost of providing service and the waiting time of a customer in
queuing system by using the probabilities and averages.
7. Replacement Models : To decide the optimal time to replace equipment, for
instance when equipment deteriorates or fails, these models are applied.
8. Dynamic Programming Model : These may be considered as an out growth
of mathematical programming, which involves optimization of multi stage,
inter-related decision process.
9. Markov Chain Models : Used for analysing a system which changes over a
period of time among various possible out comes or states.
10. Monte-Cairo Simulation : Simulation is a quantitative procedure, which
describes process by developing a model of that process and then
conducting a series of organised experiments to predict the behaviour of the
process over time. To find out how the real process would react to certain
changes, we can produce these changes in our model and simulate the
reaction of the real process to them.

L6 Significance and Applications of OR in Industrial

Problems
The Operations Research approach becomes very significant in industrial
problems particularly, when the resouces are scarce or/and when a balance is to be
brought between conflicting goals where there are many alternative courses of actions
available to the decision maker. Let us consider one such example given below.
Suppose, a decision has to be taken regarding inventory management. The
objective of production manager is based on quality of material and availability
on time. While a finance manager thinks on the line of minimising the
costs. Therefore he would prefer to order a party who can supply at lowest
costs or can offer some discount The materials manager wishes to
safeguard production from stockouts and against demand fluctuations etc.
Therefore he would like to order in such way. But the stores manager
will be particular about accommodating them in his stores and other
related problems in protecting the materials while storage and so on. Thus all
10 Operations Research : Theory and Practice

these managers are trying minimise the cost or maximise profit to the
organisation. Though every manager is thinking in the interest of organisation
growth, their objectives vary according to their specification.
In view of the situations like above a manager has to derive the decisions which
should consist.
1. A total system orientation
2. Scientific approach
3. Realistic nature
4. Models should be able to be expressed quantitatively.
5. Systematic method.
OR is the science that is embedded with suitable blend all the above features.
Thus, OR can play a significant role in bringing a balance among different
interdisciplinary people to managerial problems.
Some Areas of Applications :
1. Design of aircraft and aerospace structures for minimum weight.
2. Finding the optimal trajectories of space vehicles.
3. Design of civil engineering structures like frames, foundations, bridges, towers,
chimneys and dams for minimum cost.
4. Optimum design of linkages, cams, gears, machine tools and other mechanical
components.
5. Selection of machining conditions in metal cutting processes for minimum
production cost.
6. Design of material handling equipment like conveyors trucks and cranes for
minimum cost.
7. Design of pumps, turbines and heat transfer equipment for maximum
efficiency.
8. Optimum design of electrical machinery like motors, generators and
transformers.
9. Optimum design of electrical networks.
10. Shortest route taken by salesman visiting different cities.
11. Optimum production planning, controlling and scheduling.
12. Analysis of statistical data and building empirical models from experimental
results to obtain the accurate representation of the physical phenomenon.
13. Optimum design of chemical processing equipment and plants.
14. Design of optimum pipe line networks for process industries.
15. Selection of site for an industry.
16. Inventory control to minimise inventory costs such as ordering cost, carrying
costs, shortage costs etc.
17. Planning of maintenance and replacement of equipment to reduce the
operating costs.
18. Allocation resources of services among several activities to maximise the benefit.
19.
Operations Research - An Overview 11

Controlling the waiting and idle items and queuing in production lines to
reduce the costs.
20. Planning the best strategy to obtain maximum profit in the presence of
competitor.

L7 Characteristic features of Operations Research Model (or

Merits of Operations Research Models)
Some significant features of Operations Research Model are given below.

1.7.1 Distinct Direction in Decision Making

Operations Research model provides a clear and distinct direction to the
managers in decision making and problem solving. A major premise of Operations
Research is that decision making irrespective of the situation involved, can be
considered as a general systematic process.

1.7.2 Scientific Approach

Operations Research employs scientific reasoning to its problems. Therefore
the managers can confidently implement their decisions. Even if they fail after taking
a decision provided by Operations Research, they will have scientific and evidential
basis to plead in their support.

1.7.3 Objective Orientation

Operations Research is oriented to locate the best possible or optimal solution
to the problem. As the approach itself is embedded with setting of the goal or
objectives, it becomes easy to use this as a measure to compare the alternative courses
of action.

1.7.4 Inter - Disciplinary Team Approach

Operations Research is inter disciplinary in nature and therefore needs a team
approach. It is a blend of the aspects of various disciplines such as economics, physics,
physiology, sociology, anatomy, engineering, technology, mathematics, statistics and
management. This feature keeps Operations Research on the common berth to all
sectors of people and builds an espirit-de-corps. It can also provide a solution
acceptable to all the people,

1.7.5 Compatible to Digital Computer

Perhaps the use of digital computer has become an integral part of the
Operations Research approach to decision making. There are several software
packages developed with the Help of Operations Research approach to problems with
high volume and complexive in nature.

1.7.6 Consideration of All Factors

12 Operations Research : Theory and Practice

Operations Research takes into account of the goals (objective function) of the
organisation with all bottlenecks or hurdles (constraint set) and the feasibility
(conditions of variables). This feature provides a manager to take a decision that can
keep himself or his organisation on a competitive edge.

L8 Limitations (Demerits) of Operations Research

Operations Research, though widely used has got certain limitations. These are
given below :

L8.1 Impossibility and Infeasibility

Operations Research takes care of all the factors in choosing the best alternative.
In modern society and in real life these factors are numerous and establishing relation
among these is either impossible or infeasible in many cases. Thus many problems
will be left unsolved.

L8.2 For Intangible Situations

Operations Research can provide solution to those problems in which all the
factors can be quantifiable. But it can not provide any solution to intangible variables
and attributes which are qualitative in nature such as a human factors etc.

1.8.3 Distance Between Managers (Decision Maker) and Operation

Researcher
Operations Research is specialist's job and requires the knowledge of
mathematics, statistics etc. The researcher thus may not be aware of the business
problem and the aspect of flexibility cannot be clarified by him.

At the same time the managers may fail to understand the complex working of
the OR. Some times they may not even find time to notice any misconceptions. Thus
there is a large gap built up between one who provides solutions and one who uses
them. This leads to confusion and poses lot of problems during implementation or
practice. Therefore it may result in utter failure even with a successful formula.

Review Questions
1. Discuss the phases of Operations Research.

2. Discuss the characteristics and limitations of Operations Research.

[Mech. 96, 95/C, 97/S, 98/P, 99/CJ
Operations Research - An Overview 13

3. Discuss the types and characteristics of models used in Operations Research.

[CSE - 98/S]

4. What are the different types of models used in Operations Research?

Mention general methods of solving Operations Research models.
[Mech. 95/S, 96/S, CSE 96, 98]

5. Explain engineering applications to optimization techniques, [EEE 95, 97, 98/S]

6. "Operations Research provides bad solutions, otherwise which would have been
the worst." Comment.

7. Discuss the merits and demerits of Operations Research models.

8. Discuss the strengths and weakness of Operations Research approach.

9. Discuss the significance of OR in solving managerial problems. [CSE 2000]

10. Explain the characteristic features of OR [CSE 2000/S]

11. Write short note on 'General methods of solving OR models'

[JNTU - Mech./Prod./Chem. - 2001/S]

12. Write critical essay on the definition and scope of OR [JNTU - CSE/E.Com. E 2001]

13. Define Operations Research [0U - MBA - 90, Dec. 95, Sep 2001,]

14. State three distinguishing characteristics of OR

[MBA - 0U - May 91 Dec. 2000, Sep. 2000]

15. Explain the role of OR in management decision making [MBA - ou - March 99]

16. What is "team approach" in OR [MBA - 0U - July 2000]

17. Write any three definitions of OR [MBA - 0U - Nov. 94]

Objective Type Questions
1. Which of the following models does not use probabilities.
(a) inventory models (b) game theory models
(c) queuing models (d) linear programming

2. Which of the following belongs to Operations Research model classified on the

basis of time reference.
14 Operations Research : Theory and Practice

(a) predictive model (b) normative model

(c) dynamic model (d) simulation
3. In which of the following models, we do not use alphabet
(a) mathematical model (b) descriptive
(c) verbal (d) none of the above

4. Deterministic models are based on

(a) time frame (b) degree of certainty
(c) method of solution (d) nature or purpose of use

5. Iconic model uses

(a) scaled version (b) expectation
(c) synergy (d) time reference

6. The Operations Research model represents a system, but does not physically
resemble the components of the system is
(a) iconic (b) analogue (c) symbolic (d) normative

7. Game theory without saddle points belongs to model of OR

(a) static (b) dynamic (c) deterministic (d) probabilistic

8. Replacement models when money value not changing with time can be
considered as
(a) static (b) dynamic (c) probabilistic (d) simulative

9. Which of the following is not a strength of Operations Research.

(a) objective orientation (b) distinct decision making
(c) team approach (d) gap between the user and OR model
designer

10. Which of the following is not a limitation of Operations Research.

(a) qualitative aspects
(b) gap between decision maker and model maker
(c) involvement multiple constraints
(d) scientific approach

11. The Operations Research widely employed in marketing problems is

(a) goal programming (b) game theory

(c) dynamic programming (d) linear programming

12. "If this occurs, what will follow?" This question is connected with model.
Operations Research - An Overview 15

(a) normative (b) iconic (c) predictive (d) descriptive

13. The scaled version of a real object is

(a) iconic model (b) descriptive

(c) normative model (d) static model

14. Find odd man out.

(a) iconic model (b) verbal model
(c) mathematical model (d) probabilistic model
15. The failures of decisions based on Operations Research modeling to
managerial problems is attributed to
(a) model makers ignorance (b) managers inefficiency
(c) decision makers fault (d) understanding between decision
maker and model maker

16. In which of the following, deterministic models are not found.

(a) games and competitive strategies (b) inventories
(c) waiting line models (d) replacement models

17. Plant layout diagram is an example as best fit in

(a) probabilistic model (b) dynamic model
(c) descriptive model (d) simulative model

18. Demand forecasting can be fit as an example of

(a) static model (b) simulative model
(c) deterministic model (d) symbolic model
19. Consider these following two groups ;
1. Degree of certainty ( ) A. Iterative
2. Time reference ( ) B. Normatives
3. Application C. Stochastic
4. A method of solution D. Verbal
5. Symbolism E. Static
The correct matching is
(a) 1-C, 2-E, 3-B, 4-D, 5-A (b) 1-C, 2-E, 3-B, 4-A, 5-D
(c) A-4, B-3, C-4, D-l, E-2 (d) A-3, B-4, C-2, D-5,E-1

20. Operations Research models can not work in the case of

(a) tangible variables (b) quantitative factors
16 Operations Research : Theory and Practice

(c) human factors (d) decision maker knows OR.

21. Match the following:
1. Iconic model ( ) A. Explanation
2. Dynamic model ( ) B. Words x letters
3. Predictive model ( ) C. Changes with time
4. Stochastic model D. Expectation through clues
5. Verbal model E. Scaled version
6. Descriptive F. Comparative
G. Involvement of risk

Fill in the Blanks

1. The scaled version of a real object can be termed as model of
Operations Research.
2. Operations Research problems take into consideration of the objective
function, constraints and .
3. According to Saaty " Operations Research is choosing decision
otherwise which would have been .
4. The goodness or optimality of an Operations Research problem is checked by

5. The static and dynamic models of Operations Research are based on

reference.

Answers
Objective Type Questions :
l.(d) 2. (c) S.(d) 4. (b) 5. (a)
6.(b) 7. (d) 8. (a) 9. (d) 10. (d)
11. (b) 12 .(c) 13. (a) 14. (d) 15. (d)
16. (c) 17-<c) 18. (b) 19. (d) 20. (c)
Match the Following :
Me) 2.(c) 3.(d)
4-fe) 5.(b) 6. (a)
Fill in the Blanks :
1. iconic 2. conditions of variables 3. the bad, the worst
4. objective function 5. time
18 Operations Research : Theory and Practice

CHAPTER AT A GLANCE

Requireme
>
Exac
lity

:§v!
<

\
> u

I
*ict

a
1
5
Formulation of Linear Programming Problems 19

2.0 Introduction
In the previous chapter we have learnt that after collection of relevant data, it
is to be translated into appropriate mathematical or Operations Research model. One
of such models applied to production and allocations is linear programming in which
the variables are linearly related. This process of translation is called formulation. In
this chapter we learn how to formulate the Linear Programming Problem (LPP).

2.1 Formulation
The formulation of relevant data in Linear Programming Problem (LPP) is
carried out in the following four steps :

Step 1: Selection of decision variables.

Step 2 : Setting the objective function.

Step 3 : Identification of constraint set.

Step 4 : Writing the conditions of variables.

These are explained as follows :

Step 1: Selection of Variables :

In the given data, firstly the variables are to be identified. These are also called
decision variables or design variables.
A variable in the context of Operations Research problems is often seen in the
following ways.
(a) Number ofdifferent types ofproducts to be manufactured per day (or per week) in a manufacturing
problem (product mix).
(b) Number ofproducts to be sold in a sales problem.
(c) Number of resources (such as men f machine / material) required to do certain job in an allocation
irroblem.
(d) Number of units of certain model or type to be bought at lowest cost
(e) Number of ingredients to be mixed or used in production process problem.
if) Number of units to be ordered, in the purchasing problem
(g) Number of units to be advertised, in a marketing problem
(k) Number of units of fertilizers or ingradients in a fertilizer to be used in an agriculturing problem.
(i) Number of tablets / capsules to purchase in a medical /patients problem.
Step 2 : Setting Objective Function :
This is to set the goal in the problem. There are only two types of objective
functions in Operations Research.
These are : Maximisation (ofprofits) and Minimisation (of costs).
A maximisation type of objective function is set if the data reveals the profits in
the problem. The production, sales, revenue, output etc., are to be maximised.
20 Operations Research : Theory and Practice

A minimisation type of objective function is set if the data p r o b l e m is given in

terms of costs. T h e resources (and their utilisation) are to be minimised. T h e time
utilised at given cost, loss, wastage, shortage, spoilage, costs etc., a r e some other terms
where we think of minimisation.
Step 3: Identifying the Constraint Set: Availability (<)

A constraint set is a set of the r-+- Inequality type -

hurdles or limitations in achieving the I Requirement
objective. Constraints — constraint (>)

Equality type
T h e different types of constraints L , ,
. v . or exact type (=)
we c o m e across m linear p r o g r a m m i n g
FIGURE 2 I :
p r o b l e m are given below : TYPES OF CONSTRAINTS

1. Availability Constraint : This is related by an inequality, less than or equal to

(<) and is used when the availability of resource is limited or when maximum
limit is imposed.

Examples :
(a) Suppose you are cooking some curry dish and you have to put some salt in it, say 10 gm. You will
be constrained to put less than or equal to 10 gin-only, but excess is not allowed.
(b) Suppose you have only Rs 100/- in you wallet and went to a restaurant. You will be constrained
to order worth less than or equal to Rs. 100/- only. Here x < 100 where x is rupees spent.
(c) You have a machine on which you can utilise 8 hours a day. Thus you can have a maximum of 48
hours a week (of 6 days).
Now number of hours you can utilise on the machine (say x) < 48.

2. Requirement Constraint: This constraint is used to show a relation 'greater than

or equal to' type of inequality. This constraint is used when certain minimum
limit is imposed. Here, more than certain limit is permitted but lesser quantity
is not allowed.
Examples :
(a) You are to buy a ready madetrouser or skirt. You will be constrained to choose greater than or equal
to the required length.
(b) You have a production unit at which you have an order for 50 units. Now if the number of units
to be produced is x, then x > 50.
(c) Suppose you are pricing a product then you will say it should be greater than or equal to its
(manufacturing and transport) costs.

3. Exact Constraint: This is the constraint which can deviate to neither side. This
is represented in the equation (=)form.
Examples :
(a) You have to order the lens power for your spectacles. Here you will not permit lesser or greater than
your specified power.
(b) Wlien you purchase the shoe, you neither allow 'less than nor 'greater than' your feet size.
(c) The doctor prescribes certain quantity of antibiotic that must be the exact dosage.
(d) Certain design specification in manufacturing will be exact kind.
Formulation of Linear Programming Problems 21

Step 4 : Writing Conditions of Variables:

Here the conditions of the decision variables will be predetermined. The
conditions often found in LPP are in two kinds. These are non-negative and unrestricted.
A non-negative conditions is represented by 'greater than equal to zero' (> 0).
The above steps are illustrated with the following example.
ILLUSTRATION 1
Sainath and Co manufactures two brands of products namely Shivnath and
Harinath. Both these models have to undergo the operations on three machines
lathe, milling and grinding. Each unit of Shivnath gives a profit ofRs. 45 and
requires 2 hours on lathe, 3 hours on milling and 1 hour on grinding. Each unit
of Harinath can give a profit ofRs. 70 and requires 3, 5, and 4 hours on lathe,
milling and grinding respectively. Due to prior commitment, the use of lathe
hours are restricted to a maximum of 70 hours in a week. The operators to
operate milling machines are hired for 110 hours / week. Due to scarce
availability of skilled man power for grinding machine, the grinding hours are
limited to 100 hours/week. Formulate the data into on LPP.

-Solution :
Step 1 : Selection of Variables :
In the above problem, we can observe that the decision is to be taken on how
many products of each brand is to be manufactured. Hence the quantities of products
to be produced per week are the decision variables.
Therefore we assume that the number of units of product Shivnath brand
produced per week = Xy.
The number of units of product of Harinath brand produced per week = x2.
Step 2 : Setting Objective :
In the given problem the profits on the brands are given.
Therefore objective function is to maximise the profits.
Now, the profit on each unit of Shivnath brand =Rs. 45.
Number of units of Shivnath to be manufactured = Xy
.*. The profit on xl units of Shivn-ath brand = 45 xY
Similarly, the profit on each unit of Harinath brand = Rs. 70
Number of units of Harinath brand to be manufactured = x2
.*. The profit on x2 units of Harinath brand = 70 x 2
The total profit on both brands = 4 5 ^ + 70x2
This total profit (say z) is to be maximised.
Hence, the objective function is to
Maximise z = 45xj + 70x2
22 Operations Research : Theory and Practice

Step 3 : Identification of Constraint Set:

In the above problem, the constraints are the availability of machine hours.

1. Constraint on Lathe Machine : Each unit of Shivnath brand requires 2 hours/

week
So xx units of Shivnath brand requires 2xx hours / week.
Each unit of Harinath brand requires 3 hours / week
and So x2 units of Harinath brand require 3x2 hours / week.
Total lathe hours utilised for both the brands is 2x1 + 3x2
and this cannot exceed 70 hours/ week.

(Constraint on availability of lathe hours due to prior commitment)

2. Constraint on Milling Machine : Milling hours required for each unit of Shivnath
brand = 3 hours/week.

/. For xj units = 3xj hours/week

Milling hours required for each unit of Harinath brand = 5 hours/week.
.*. For x2 units = 5x9
Total milling hours = 3x± + 5x2
This can not be more than 110

(Constraint on availability of milling machine hours due to hiring)

5. Constraint on Grinding Machine :

One unit of Shivnath needs one hour/week and xx units needxj hours/week
One unit of Harinath needs 4 hours/week and x2 units need 4#2 hours/week.
Total grinding hours = xj + 4x2 and this cannot be greater than 100 hours.
100
(Constraint on availability of grinding hours due to scarcity of skilled labour)

Step 4 : Writing Conditions of Variables :

Bothxj and x2 are the number of products to be produced. There can not exist
any negative production. Therefore xx and x2 can not assume any negative values (i.e.,
non negative)
Mathematically
xx > 0 and x2 > 0
Formulation of Linear Programming Problems 23

Step 5 : Summary :
Maximise / Z = 4 5 x 1 + 70x2
Subject to £>V l i l y < ^70
J^j _|_ ^ ^ i _ < nio
X:j +4x 2 < 100
X;j > 0 and x 2 > 0
Some More Examples :
ILLUSTRATION 2
Formulate thefollowing problem as an LP Problem;
Afirm engagedin producing 2 models Xv X2performs 3 operations Painting.
Assembly and Testing. The relevant data are as follows:

Hours Required for each unit

Model Unit Sales Price
Assembly Painting Testing
Model X2 Rs. 50 1.0 0.2 0.0
Model X2 Rs. 80 1.5 0.2 0.1

Total number of hours available are:

For Assembly 600
ForPainting 100
For Testing 30
Determine weekly production schedule to maximise revenue. [ JNTU Mech. 98/8]
Solution ;
Note : Only Formulation part of the solution is given here

Variables :
Let the weekly produced units of model Xi = x^
i.e., Let the no. of units produced per week in Model Xj — x^ and let the no. of
units produced per week in model X2 = ^ 2
Objective Functio :
Unit sales price for model Xy — Rs. 50/- .
For Xy units, weekly sales revenue = 50 x^.
Similarly, unit sales price for model X2 = Rs. 80/-.
For x2 units, weekly sales revenue = 80 x2.
Total sales revenue per week = 50 x} + 80 x2.
Hence the objective function is to -
Maximise Z= 50 x± + 80 x2
24 Operations Research : Theory and Practice

Constraint Set:
For assembly,
one unit of model X1 requires 1.0 hrs.
.-. xj units require 1.0 xj hrs.
one unt of model X2 requires 1.5 hrs.
x2 units of model X2 requires 1.5 x2 hrs.
Total assembly hours per week = 1.0 x^ + 1.5 x2.
This cannot exceed 600 hrs since only 600 hrs are available for assembly.
1.0*2+
(Constraint on availability of assembly hours)
Similarly,
For painting
xy units require 0.2 xy hrs
and x2 units require 0.2 x% hrs
0.2 xj + 0.2 x2 < 100
(Constraint on availability of painting hours)
For Testing

or
0.1 x% < 30
(Availability constraint on testing hours)
Conditions of Variables :
As the problem is to determine weekly production schedule and we cannot have
a negative production.
Therefore,
xx > 0 and x2 > 0
Summary of Foramulation :
maximize Z = 50 xy + 80 x2
subject to 1.0 x1 + 1.5*2 -
0.2^! + 0.2x 2 < 100
0.1x 2 <30
xy , x2 > 0
Formulation of Linear Programming Problems 25

ILLUSTRATION 3
A firm manufactures two products in three departments. Product A contributes
Rs, 5/- unit and requires 5 hrs, in dept, M9 5 hrs. in dept, Nand one hour in dept, P,
Product B contributes Rs. 10/- unit and requires 8 hrs, in dept My 3 hrs in dept N
and 8 hrs in dept R Capacities for departments M9 N, P are 48 hours per week.
Find out optimalproduct mix using Simplex model [JNTUCSE 98]

S o l u i t o n •;
Note : Formulation part of the solution is given here, for solution by simplex ref. illustration 18, chapter 4

The problem is summarised below.

A B Capacity
M 5 8 48
N 5 3 48
P 1 8 48
Unit contribution 5 10
Variables :
Let quantity produced of product A = Xj
Let quantity produced of product B - X2
Objective Function :
Profit contribution of each unit of A = Rs, 5
Profit contribution of Xx units of A = 5 Xj
Profit contribution of each unit of B = Rs. 10
Profit contribution of X2 units of B = 10X2
Total profit = S X ^ l O X g
Therefore objective function is to maximise Z = 5Xj + 10 X2
Constraint Set:
In Department M
Time required by each unit of product A = 5 hrs
Time required by Xt units of product^ = 5 Xj
Time required by each unit of product B = 8
Time required by X2 units of product B = 8 X2
Total hours available in dept. M = 48 hrs per week
.\ I 5 Xx + 8 X2 < 48 I
(Constraint on capacity in dept. M)
Similarly in department AT,
5 Xj + 3 X2 < 48
(Constraint on capacity in dept. N.)
26 Operations Research : Theory and Practice

and in department P,

(Constraint on capacity in department P)

Conditions :
Xx and X2 > 0

Summary :
Max 2 = 5 X 1 + 10X2
Subject to 5 Xj + 8 X2 < 48
5 Xx + 3 X2 < 48
Xx + 8 X2 < 48
xx, x 2 > o
ILLUSTRATION 4

Food X contains 6 units of VitaminsAper gram and 7 units of Vitamin B per gram
and costs 12 paise per gram. Food Ycontains 8 units of Vitamin Aper gram and 12
units of Vitamin B and costs 20 paiseper gram. The daily minimum requirements
of Vitamin A and Vitamin B are 100 units and 120 units respectively. Find the
minimum cost of product mix. Use simplex method. [ JNTU B.Tech. (ECE) 97]

Solution i

Note ; Formulation part of the solution is given here, refer illustration - 7, chapter 4 for its
solution by simplex method
Selection of Variables :
Let the No. of grams produced in type X = xy
Let the No. of grams produced in type F= x2
Objective Function :
Cost of each gram of type X= 12 paise
Cost of xy gms of type X- 12x1 ps.
Cost of each gm of type F=20 ps.
Cost of x2 gms of type F=20 x2 ps.
Total cost Z = 12^ + 20*2
.-. Objective function is to minimize Z= 12#i + 20#2
Constraint Set:
For Vitamin A :
Availability of vitamin y4 in each gm of food X = 6 units
Availability of vit-^4 in x± gms of food X = 6 xx
Availability of vit-^4 in each gms of food F = 8 units
Formulation of Linear Programming Problems 27

Availability of vit-^4 in x2 gms of food Y = 8 x2

Total vitamin/I = 6x1 + Sx2
Daily minimum vitamin^ required =100 units
> 100
(Requirement constraint on vitamin A)
For Vitamin B :
Availability of vitamin B in each gm of food X = 7 units
Availability of vit-B in x^ gms of food X=7 x±
Availability of vit-B in each gms of food Y= 12 units
Availability of viuB in x2 gms of food F = 12 x2
Total vitamin 5 = 7 ^j + 12 x2
Daily minimum requirement of vitamin B = 120 units
Hence 12 x2 > 120
(Constraint on requirement of vitamin — B).
Conditions of Variables :
xx and x2 cannot be negative
*!>0 8c x2>0

Summary :
min z= 12 x^ + 20 x2
subject to 6:^ + 8 # 2 - 100
12x 2 > 120
( 0
>
ILLUSTRATION 5
MA ABCL company manufactures two types of cassettes, a video and audio. Each
video cassette takes twice as long to produce one audio cassette, and the company
would have time to make a maximum of 2000 per day if it is produced only audio
cassettes. The supply of plastic is sufficient to produce 1500 per day to both audio
and video cassettes combined. The video cassette requires a special testing and
processing of which there are only 6000 hrs. per day available. If the company
makes a profit ofRs. 3/- and Rs. 5/-per audio and video cassette respectively, how
many of each should be produced per day in order to maximize the profit?
[JNTU2003(Set-l)]

Let No of audio cassettes to be produced per day - x±

No. of video casettes to be produced per day = x2
Objective Function :
Profit per one audio casette = Rs. 3
Profit on xx audio casettes = 3x
Profit on each video casette = 5
28 Operations Research : Theory and Practice

Profit on KJ video casettes = 5x2

Total profit = 3 x± + 5 x2

It is to be maximized
.-. Objective function is to Maximize
Z = 3xj + 5x2

Constraint Set:

(i) Time Constraint :

As time of production of video is twice to that of audio cassettes, if xx audio are

x2
produced — videos can be produced. Thus in terms of audio casette times, we get

X
2
x± + — < 2000 (constraint on production time)

(ii) Plastic Constraint :

xx+x2 < 1500

(Hi) Testing & Processing Time Constraint :

x2 < 6000

Conditions Both x± > 0 and x9 > 0

Summary Maximize Z = 3xj + 5x2

subject to 2xx + x2 < 4000

xx +x2 < 1500

x2 < 6000

x1? x2 > 0
Formulation of Linear Programming Problems 29

ILLUSTRATION 6

A manufacturing firm producing refrigerators has given a contract to supply

50 units at the end of the first month, 50 at the end of the second month, and 50 at
the end of the third. The cost of producing fXfno. of refrigerators in any month is
given by fX2\ The firm can produce more number of refrigerators in any month
and carry them to the next month. However, a holding cost ofRs. 20per unit is
charged for any refrigerator carried over from month to the next Assuming
there is no initial inventory, determine the no, of refrigerators to be produced in
each month so as to minimize the total cost, using dynamic programming
approach. [JNTU B.Tech (ECE) 97/S]

Solution :
Note : Formulation part is discussed here. For complete solution refer chapter 12, illustration 6

Selection of Variables :
Let, x{ = No. of refrigerators to be manufactured in first month
=
x^ No. of refrigerators to be manufactured in second month

#3 = No. of refrigerators to be manufactured in third month.

Objective Function :

Now, Given the production cost of X units manufactured = X2,

we have, Production cost in first month = x\

Production cost in second month = x\

and Production cost in third month = x\

and there is no initial inventory, hence no holding in first month.

But, the holding cost in second month if xj exceeds 50,

= 20 (*! - 50)

also, the holding cost in third month if x1 + x2 exceeds 100 is,

= 20(x 1 +x 2 -100)

Hence, the total cost

= x\ + x\ + x\ + 20 (*! -50) + 20 (xY +x2- 100)

Objective function is to

Minimize Z = x\ + x\ + x\ + 40 xx + 20 x2 - 3000
30 Operations Research : Theory and Practice

Constraint Set:

According to contracts of supply,

Xi > 50 (Constraint on supply for first month)

Xi + #2 > 100 (Constraint on supply for first two months)

Assumed that no stock is to be left out after third month.

#1 + #2 + #3 — 150 (Constraint on supply for the three months)

Condition of Variables :

All n o n negative i.e., x^ , ac2 , #3 > 0.

Summary :

Minimise z = x\ + x\ + x\ + 40 x^ + 20 x2 - 3000

subject to xj > 50

xx+x2> 100

Xj , JC9 , x 3 > 0.

ILLUSTRATION 7

A company wants to purchase at most 180 units of a product There are two types
of the product, M} andM2 available. M} occupies 2fi> cost $ 12 and the company
makes a profit of $3. M3 occupies 3fi, cost $15 and the company makes a profit of
$4. If the budget is $15,000 and the warehouse has 3000ftforproduct
Solve the problem using Simplex method.
[JNTU B.Tech. (EEE) 98]

Solution :

Note : Solution for (i) is discussed here

Selection of Variables :

Let the no. of products purchased in type Mj =

and the no. of products pruchased in type M 2 =

Formulation of Linear Programming Problems 31

Objective Function :
Profit on each unit of product of type Mj = $ 3

Profit on xj units of product of type My = $ 3 xj

Profit on each unit of product of type M2 = $> 4

Profit on x2 units of product of type M 2 = $ 4 x%

Total profit Z = 3xj + 4xy

.-. Objective function is maximize Z = 3 xi + 4 #2

Constraint Set:
Space Constraint ;

Each unit of product Mi occupies 2 /£3 space.

.-. xy units of product M± occupy 2 ^i/^ 3

Similarly, each unit of product M 2 occupies 3 ft^

x2 units of product M 2 occupy 3#2/£3

Total space for both products 2 ^ + 3x2

Available space = 3000/^3

2 Xi + 3^2 < 3000 (Constraint on availability of space)

Budget Constraint:
Each unit of product My costs $ 12

x1 units of product M^ cost $ 12 ^j

and each unit of product M 2 cost $ 15

x% units of product M 2 cost $ 15 x%

Total cost = 12 xi + 15 x2

Available budget = $ 15, 000

12 #1 + 15 %2 - 15000 (Constraint on availability of budget)
32 Operations Research : Theory and Practice

Condition of Variables :
both xi and x2 can not be negative
.. | x 7 > 0 and x2 > 0 [
Summary :
Maximise Z = 3xj + 4x2
subject to
2x{ + 3x2 ^ 3000

15x 2 < 15000

ILLUSTRATION 8

,4w oil refinery can blend three grades ofcrude oil to produce quality P and Qpetrol.
Two possible blending processes are available. For each production run, the older
process uses 5 units of crude A, 7 units of crude Band 2 units of crude C to produce
9 units ofP and 7 units ofQ. The newer process uses 3 units of crude A9 9 units ofB
and 4 units of crude C to produce 5 units ofP and 9 units ofQ petrol. Because of
prior contract commitments the refinery mustproduce at least 300 units ofP and at
least 300 units of Q for the next month. It has available 1500 units of crude A,
1900 units of crude B and 1000 units of crude C For each unit ofP the refinery
receives Rs. 60 while for each unit of Q it receives Rs. 90. Find out the linear
programming formulation so as to maximise the revenue.

Selection of Decision Variables :

x\ = No. of production runs of the older process
x2 = No. of production runs of the newer process.
Objective Function :
No. of units of petrol P in older process-for each production run = 9
No. of units of petrol P in older process for x^ production runs = 9 xj
No. of units of petrol P in newer process for each production runs = 5
No. of units of petrol P in newer process for x2 production runs = 5x2
Total petrol P = 9*i + 5x2
Profit on petrol P = 60 (9xj + 5x2 ).
Similarly, for petrol Q
No. of units per production run in older process = 7
No. of units for x} production runs = 7 xl.
Formulation of Linear Programming Problems 33

No. of units for each production run in newer process = 9

No. of units for x2 production runs in newer process = 9 x2

No. of units of petrol Q = 7 x± + 9x2

Profit on petrol g = 90 (7xx + 9x2 )

Total profit for both P and Q = 60 ( 9 ^ + 5x2 ) + 90 (7xl + 9x2 )
Objective function is
Maximize Z = 1170 xx + 1110 x%.

Constraint Set:
For Petrol P :
No. of units of petrol produced in both processes is 9xj + 5x2. This should be
at least 500 units.
.-. 9xj + 5x 2 > 500
(Requirement constraint on contract commitment ofP)
For Petrol Q:
No. of units of petrol produced in both processes is 7#j + 9x2- This should be
at least 300 units.
.-. 7xx + 9*2 > 300
(Requirement constraint on constract commitment ofQ).
For Crude A :
For older process, each production run requires 5 units.
For older process, X} production runs require 5x1 units.
For newer process, each production run requires 3 units.
For newer process, x2 production runs require 3x2 units.
Total crude A for both processes 5x1 + 3x2.
This can not be more than 1500.
.. 5xx + Sx2 < 1500
(Constriant on availability of crude A)
For Crude B :
Similarly, 7xx + 9x2 < 1900
(Availability constraint on curde B).
For Crude C :
Similarly, 2xx + 4JC2 £ 1000
(Availability constraint on crude C)
34 Operations Research : Theory and Practice

Conditions :
No. of production runs can not be negative.
Hence *x > 0 and * 2 > 0
Summary :
Maximise Z = 1170 x} + 1110 x2
Subject to 9xx + 5x2 ^ 500
7xj + 9x2 > 300
5*j + 3* 2 < 1500
lxx + 9x2 < 1900
2xx + 4x2 < 1000
xj , x2 > 0
ILLUSTRATION 9

An advertising company has to plan their advertising strategy through the

different media viz, TV, radio and newspaper. The purpose of advertising is to
reach maximum number of potential customers. The cost of an advertisement in
TV, Radio and newspaper are Rs. 3,000/-; Rs. 2000/- andRs. 2500 respectively.
The average expected potential customers reached per unit by 20000 of which
15000 are female customers. These figures with radio are 60000 and 40000 and
with newspaper 25000 and 12000 respectively. The company has a maximum
budgetfor advertising as Rs. 50000/- only. It is proposed to advertise through TV
or radio between 6 and 10 units and at least 5 advertisements should appear in
newspaper. Further, it decides that at least 1,00,000 exposures should take place
among female customers. Budget of advertising newspapers is limited to
Rs. 25000 only. Formulate into LPR

Decision Variables :
Number of advertisements by TV = Xj
Number of advertisements by Radio = x2
Number of advertisements by Newspaper = x3
Objective Function :
To maximise potential customers.
Max. 2 = 20000*! + 60000*2 + 25000*3
Constraint Set:
(i) Budget Constraint:
3000xj + 2000x2 + 2500x3 < 50000
or 30x| + 2Ox:2 + 25x3 < 500
or 6xx + 4x2 + 5*3 < 100
(ii) Female Customer Constraint:
15000xj + 4000(k2 + 12000*3 < 100000
or 15*! + 40*2 + 12
*3 - 10
°
Formulation of Linear Programming Problems 35

(Hi) Cost Constraint on Advertisements in News Paper :

2500x3 < 25000
or xs < 10
(iv) Constraint on Number ofAdvertisement in News Paper :
x3>5
(v) Constraint on Number ofAdvertisements in TV:
6<xx< 10
or xj > 6, xx < 10
(vi) Constraint on number of advertisement in radio :
6<x2< 10
or x2 > 6 and x2 < 10
Conditions :
All the variables are non negative, since there will be no negative number of
exposures

Summary :
Max Z = 20000*! + 60000x2 + 25000*3
Subject to 6xy + 4x2 + 5x3 < 100
15^+40x9+ 12x3< 100
xs < 10; x$ > 5
x{>6; Xi<10

x2> 6; x 2 < 10
and Xy, x2y x% > 0

ILLUSTRATION 10
MK Raju has received his retirement benefits such as provident fund, gratuity
etc. He wants to invest this money in Government securities, term deposits, in
banks, investment on company deposits, equity shares and house construction.
The return on investment (ROT), the time limits and risk on afivepoint scale is
given in thefollowing tabularform.
No. of
S.No. Type of Investment Return Risk
Years
1. Govt. Securities 8% 7 1
2. Company deposits 15% 4 3
3. Term deposits 14% 3 2
4. Equity shares 18% 6 5
5. House construction 25% 10 1
He decides that the average risk should not exceed 3 and his funds should not be
locked upfor more than 10 years. He also has necessarily to invest 30% on house
construction. Help Mr. Raju to plan his investment so as to maximise his ROL
36 Operations Research : Theory and Practice

Solution ;
Decision Variables :
Let the percentages of investments on these five factors be Xj, x2 . . . x$
Objective Function :
Maximise Z = 8 ^ + 15#2 + 14x3 + 18x 4 + 25* 5
Subject to constraints
7xx + 4x2 + 3x 3 + 6tf4 + l(k 5 < 10
(Constraint on term of investvient)
xy + 3x2 + 2x$ + 5# 4 + x5 < 3

(Constraint on risk)
x5 > 0.3
(Constraint on investvient on house construction)

(Total percentage is 100)

Conditions of Variables :
X\, x%, #3, x4, x5>0

ILLUSTRATION 11
A pregnant woman is advised to take Iron, Zinc and Folic acid, at least in the
quantities 100 mg, 150 mg and 100 mgper day respectively. There are two types of
medicines available in the form of a 500 mg tablet and a 250 mg capsule. Each
tablet is composed of30 mg of Iron, 45 mg of Zinc and 40 mg ofFolic acid while a
capsule contains the three ingradients as 20,15, 25 mg respectively. The cost of
each tablet is Rs. 3/- while that ofeach capsule is Rs. 2/-. Determine the number of
tablets and capsules to be purchased so that the total cost of medicines is
minimized.
Solution :
Selection of Variables:
Let, Number of tablets of 500 mg = x1
Number of capsules of 250 mg = x2
Objective Function :
Each tablet costs Rs. 3/-, xx tablets will cost 3x{
Each capsule costs Rs. 2/-, x% capsules will cost 2x2
Total cost = 3x + 2x2
.*. Objective function is to
Minimize Z = 3xj + 2x%
Formulation of Linear Programming Problems 37

Constraints :
Iron Constraint :
Iron in each tablet - 30 mg
Iron content in xj tablets = SOx^
Iron content in each capsule = 20 mg
Iron in x2 capsules = 20 x2
Total Iron = 3 0 ^ + 20x9
,\ 30xT + 20x 2 > 100
or 3x1 + 2^2 ^ 10
(Constraint on daily viin. requirement of iron)
Similarly, the constraints on zinc and folic acid requirements can be written as
4 5 ^ + 15x2 > 150
or 3*! + x2 > 10
and 40bcj + 25x2 > 100
or &CJ + 5x2 > 20
Conditions :
x±,x2 being number of tablets capsules can not be negative (i.e., xl9x%> 0)
Summary :
Minimise Z= 3xj + 2x9
Subject to 3xj + 2x2 > 10
3x y + X2 > 10
Sxj + bx2 > 20
x 1 ,x 2 > 0.
ILLUSTRATIOW 12
Medison Nursing Home works 24 hours a day. In a ward, it requires the attendance
of nurses as given below:
From To Number of Nurses required
00.00 04.00 2
04.00 08.00 3
08.00 12.00 5
12.00 16.00 4
16.00 20.00 7
20.00 0.00 3
Consider that any nurse can work only for 8 hours in a day. Formulate the
information in an LPPso as to schedule with minimum manpower.
(JNTUB.Tech Mech. & Met. 95/S)

Decision Variables :
X! = Number of nurses added at 00.00
x2 = Number of nurses added at 04.00
x s = Number of nurses added at 08.00
38 Operations Research : Theory and Practice

x4 = Number of nurses added at 12.00

x5 = Number of nurses added at 16.00
XQ = Number of nurses added at 20.00
Objective Function :
The objective is to minimise the man power requirement. Therefore we must
minimise the total number of nurses required in all the periods.
Min. Z = %i + %2 + #3 + x4 + #5 + x6
Constraints :
At any four hour period, the number of nurses added and continuing from the
previous shift should be at least equal to their requirement in the corresponding
period.
Also, the nurses have to work for eight hours and the periods are split in 4 hour
periods, any two consecutive periods can be clubbed to arrange the number of
required nurses. Thus we get,
Constraint on period 00.00 to 04.00
Number of nurses continuing from previous period (i.e., added at 20.00) = x6
Number of nurses added at 00.00 = xx
Total number of nurses = x§ + xj and this should be at least 2 in 00.00 to 04.00.
Therefore x§ + x± > 2 (Constraint on requirement from 0.00 to 04.00)
Similarly x1 + x 2 ^ 3 (Constraint on requirement from 04.00 to 08.00)
x% + x$>5 (Constraint on requirement from 08.00 to 12.00)
X3 + X4 > 4 (Constraint on requirement from 12.00 to 16.00)
x4 + x5 > 7 (Constraint on requirement from 16.00 to 20.00)
#5 + x$ > 3 (Constraint on requirement from 20.00 to 0.00)
Conditions :
Since the no. of nurses can not be negative and cannot be a fraction value, we
write the constraints as
x- > 0 and integral value or x- is positive integer where i- 1,2, 6
ILLUSTRATION 13 —
Afarmer has 100 acrefarm and can sell all tomatoes, lettuce or radishes that he can
raise in the farm. The prices of these are Rs. LOO per kg, 0.75 per head andRs. 2/-
per kg respectively. The average yield per acre is 2,000 kg of tomatoes, 3000 heads
of lettuce and 1000 of radishes. Fertilizer is available at Rs. 0.50 per kg and the
amount required per acre is 100 kg each for tomatoes and lettuce and 50 kg for
radishes. Labour required to sow, cultivate and harvest per acre is 5 man days for
tomatoes and radishes and 6 man daysfor lettuce. A total of400 man days of labour
atRs. 20/-per man day is available.
Optimise the totalprofit of the farmer.
Formulation of Linear Programming Problems 39

S o I uii o KV : (Only formulation is shown here).

Decision Variables :
Let number of acres allocated to tomatoes = x±

Number of acres allocated to lettuce =.x2

Number of acres allocated to radish = x3

Objective Function :
Here, we have to maximize the farmer's profits which is sales minus costs.
Sales:
Land for tomatoes -Xy acres.

Yield of tomatoes = 2 OOOx x

Tomatoes total sale = 2000x x x 1.00

= 2000*!

Similarly, sales of lettuce = 0.75 x 3000 x2

and sales of radish = 2.00 x 1000%3

Total sales = 2OOOx2 + 2250x2 + 2000x3

Labour Cost:
Man days needed for tomatoes and radishes = 5 per acre
Land for tomatoes and radishes - x± + x 3 acres.

Total man days = 5 (#j + #3)

Man days needed for lettuce = 6 per acre

Land for lettuce = x2 acres

.-. Man-days for lettuce = 6x2

Total man days = 5xj + 6^2 + 5x3

Unit labour cost = Rs. 20 per man-day
.-. Total labour cost = 20 (5xj + 6x2 + 5x3)
= lOQoc^ 120x2+

Fertilizer Cost:
Requirement for tomatoes and lettuce = 1 0 0 % / acre
Land of tomatoes and lettuce = xx + x2
.-. Fertilizer required for tomatoes =100 {x^ + x2)
Requirement for radish = 50 kg/ acre
40 Operations Research : Theory and Practice

Land of radish = x3 acres

Fertilizer needed for radish = 5Gx:3
Total fertilizer = 1 0 0 ^ +100x2 + 50%3 kg
Cost of fertilizer = Rs. 0.50 per kg
Total cost on fertilizer = 0.50 (lOOxj + 100c2 + 50x3)

Now profit = sales - costs

= sales revenue - [labour cost + fertilizer cost]
= (2000^ + 2250x2 + 2000x3)
- 1(100*! + 120x2 + 10Qx:3)

The objective function is to

Maximise Z = 1850*! + 208QJC2 + 1875K 3

Constraints :
Land Constraint:
Total land used = Xj + x2 + #3
Land available =100 acres
.-. xx + %2 + ^3 < 100
(Constraint on availability of land)
Labour Constraint :
Total labour days required = 5x^ + 6x2 + 5x3 man-days
Available labour days = 400 man-days
.-. 5xx + 6x2 + 5#3 < 400
(Constraint on availability of man-days)
Conditions :
Since land cannot be negative
x\9x%9xs>0
Summary :
Max Z = 1850 xx + 2080x2 + 1875x3
Subject to Xj + x^ + x3 < 100

xl > 0, x2 < 0,
Formulation of Linear Programming Problems 41

ILLUSTRATION 14
Old hens can be bought atRs. 20 each and young ones atRs. 50 each. The old hens
lay 3 eggs per week and the young ones lay 5 eggs per week, each egg being worth of
Rs. 1.50ps. A hen (young or old) costs Rs. 1.50 per week to feed, I have only Rs. 800
to spend for hens, how many of each kind should I buy to give a profit of at least
Rs. 60/-per week, assuming that I cannot house more than 2ffhens.
[JNTU B.Tech. Mech.]
S o I ufi cm :
Decision Variables :
Number of old hens = xx
Number of young hens = x2
Objective Function :
Number of eggs each old hen lays =3
Number of eggs laid by x^ old hens = 3xl
Number of eggs each young hen lays =5
Number of eggs that x2 young hens lay = 5x2
Total eggs = Sx j + 5x2
Price of each egg =1.50
Sales revenue on eggs =1.5 (3*^ + 5x2)
= 4.5x1 + 7.5x2
Total expenditure for feeding (*j + *2)
hens @ Rs. 1.50 per week = 1.5 (xy + x2)
Profit = sales - expenditure
= (4.5*! + 7.5x2) - (1.5*! + 1.5*2)

Hence, the objective function is to Maximise Z = 3*^ + 6x2

Constraints :
Profit Constraint:
The profit should be at least Rs. 60
3*j_ + 6x 2 i>60
Constraint of Money:
Cost of old hen = Rs. 20
Cost of *! old hens = 20^
Cost of each young hen = Rs. 50
Cost of x2 young hens = 50x9
Total cost on hens = 20*^ + 50x2
Since I have Rs. 800/- only,
20*!+50*2 ^800 or 2xx + 5*2 < 80
(Constraint on availability ofvioney)
42 Operations Research : Theory and Practice

Constraint of Space:
Total number of hens = xy + x2
Available space is sufficient for 20 hens
xx + x% < 20
(Constraint on availability of space)
Conditions :
It is not possible to purchase negative quantity of hens
Therefore, *j > 0, x% > 0
Summary :
Maximise Z = 3 ^ + 6x2
Subject to 3x± + 6x2 > 60
2xY + 5x2 < 80
Xj+^^20
and xj > 0,x2 > 0.

2,2 Major Assumption ofLPP

The following are the assumptions (Charateristics) of LPP.

1. Proportionality: A primary requirement of LPP is that the obj ective function

and all constraints, are linear and hence these are directly proportional.
Linearity implies that the product of variables (likex jx 2 ), powers of variables
(like x\,x\ etc.), combinations (e.g. ax j + b logx 2 ) and exponentials (e.g.
ex , ax etc.) are not allowed.

2. Additivity : The LPP provide concept of additivity in the quantities having

same units. For example a machine needs t^Sct^ hrs. for product A and B
respectively then total time can be said to ty +19 for both products on the
machine. However, this additivity does not hold good for the quantities of
different chemical composition.

3. Multiplicativity : LPP provides proportionality and thus if one item costs

Rs. 10, then 10 items will cost Rs. 100/-.
4. Divisibility : LPP permits fractional values in addition to the integral values.
5. Deterministic : All parameters in LPP are assumed to be known exactly.
However actual production may depend on chance also.
6. Sensitivity : Further, the problem can be extended by sensitivity analysis to
check the post optimal situations.
7. Decision Making by Conditions : The conditions on the answers stated at the
beginning of the problem will aid the decision making. The assumptions
such as non negativity or unrestricted variables can influence the answers.
Formulation of Linear Programming Problems 43

Practice Problems
1. A mining company is taking a certain kind of ore from two mines X and K The
ore is divided into 3 quality groups A, B and C. Every week the company has to
supply 240 tonnes of A, 160 tonnes of B and 440 tonnes of C. The cost per day
for running the mine X is Rs. 3000, while it is Rs. 2000 for the mine Y. Each
day, Xwill produce 60 tonnes of A, 20 tonnes of B and 40 tonnes of C. The
corresponding figures for Fare 20,20 and 80.

Develop the most economical production plan by finding the number of days
for which the mines X and Y should work per week. [June 2003 (Set-2)]
Answer : cost = Rs. 18000/-
2. A person requires 10, 12 and 12 units of chemicals^, B and C respectively for
his gardens. Aliquid product contains 5, 2 and 1 units of/4, B and C respectively
per jar. A dry product contains 1, 2 and 4 units of A, B and C per carton. If the
liquid product sells for Rs. 3 per jar and the dry product sells for Rs. 2 per
carton, how many of each should be purchased to minimize the cost and meet
the requirements? Formulate the above problem as a LPP and solve it by
graphical method. [June 2003 (Set-3)]
Answer : cost = Z m j n = Rs. 13
3. A company has two bottling plants, one located at Bangalore and the other at
Mysore. Each plant produces three brands of drinks A, B and C. Bangalore
plant can produce (in one day) 1500, 3000 and 2000 bottles of A, B and C
respectively. The capacity of Mysore plant remains, 1500, 1000 and 5000 bottles
per day of A, B and C respectively. A market survey indicates that during the
month of April there will be demand of 20,000 bottles of A, 40,000 bottles of
B and 44,000 bottles of C. the operating cost per day for the plant at Bangalore
is Rs. 600/- while the operating cost per day for the plant at Mysore is Rs. 400/-
For how many days each plant be run in April so as to minimize the production
cost while still meeting the demand? [June 2003 (Set-4)]
Answer: cost of operation is Rs. 8800/-
4. Hema Pens Ltd. produces two types of pens namely; Supermo(S) and Economy
pens (E). The net profits on these types are Rs. 10/- and Rs. 6/- respectively.
Raw material required for Supermo(S) is twice as that of Economy (E). The
supply of raw materials is sufficient only for 1000 pens of E per day. Pen S
requires a special nib and Pen E requires ordinary nib. There are 400 special
nibs, 700 ordinary nibs available in stock. Formulate for the daily product mix
so as to maximise the total profit of company. [Dr. BRAOU MBA - 2004 Assign.]
Answer : Max Z = 10 x j -+ 6 x% ;
xx
S.t. — + x9 :

x}<400;
x2 < 700
,#2-0 and integers
44 Operations Research : Theory and Practice

5. Sreeja & Co wishes to plan its advertising strategy. There are two media under
consideration, Siti cable and Popular channel. Siti cable has a reach of 2000
potential customers and Popular channel has a reach of 3000 potential
customers. The cost per appearance of one minute is Rs. 6000 and Rs. 9000 in
Siti and Popular respectively. The budget of Sreeja is Rs. 80,000 per month.
There is an important requirement that the total reach for the income group
under Rs. 60,000 per annum should not exceed 3000 potential customers. The
reach in Siti cable and Popular channel for this income group is 300 and 150
potential customers. How many appearances of one minute advertisements
should Sreeja plan so as to maximise the total reach?
Answer: Max Z = 2000 xy + 3000x 2
S.t. 6000 xx + 9000 x2 < 80000
300 xj + 150 x2< 3000
x±, x2 > 0

6. Himaja Electrical Equipments (P) Ltd is engaged in the production of high

power and low power transformers using three operations : mould making, coil
assembly and joining 8c final assembly. A high power transformer is sold at Rs.
80,000 while low power transformer is sold at Rs. 3,000/-. The costs of HPT and
LPT are Rs. 3000 and Rs. 2000 respectively. The time required in hours in each
operation, and their weekly availability are given below.
Transformer
Operation Weekly Availability
High - Power Low - Power
Mould making 3 hours 1 hours 40
Coil assembly 16 3 150
Joining & final assembly 6 2 100
Formulate the problem into an LPP so as to maximise the profits.
Answer : Profit = sales price - cost
Max Z = 5000 xY+ 1000 %2
S.t. 3*! +# 2 < 4 0 ; 16x| + 3x 2 < 150
100 ; xl9x2>0
Note : Third constraint is redundant since it is already covered by first constraint
7. The manager of an oil refinery has to decide on the optimal mix of two possible
blending processes of which the inputs and outputs per production run are
given as follows:
Process Inputs Out puts
Crude A Crude B Gasoline X Gasoline Y
I 5 3 5 8
II 4 5 4 4
The maximum amount available of crude^4 and B are 200 units and 150 units
respectively. Market requirements show that at least 100 units of petrol X and
80 units of petrol Y must be produced. The profit per production run from
process I and process II are Rs. 300 and Rs. 400 respectively. Formulate the
problem as a linear programming model.
[0U - 87, 0U - MBA 90, Sep. 2001] [JNTU Mech - 95]
Formulation of Linear Programming Problems 45

Answer : x^ x2 = No. of units from Process I & II respectively.

MaxZ=300x 1
Subj. to 5xj
1 2 2 h x2 > 0
8. MIDHANI produces an alloy of specific gravity 0.98, chromium content not
more than 8% and melting point below 450° C. Three raw materials Mv M2
andM 3 are used to manufacture this alloy. The properties of these materials is
as follows :
Property of R/M
Property
Mx M2 M3
Specific gravity 0.92 0.97 1.04
Chromium content 7% 13% 16%
Melting point 440° C 490° C 480° C

Costs of the raw materials are Rs. 90, Rs. 100 and Rs. 80. Formulate to
determine the proportions that will minimise the total costs.
Answer: Percentage of RM's are x1? x%, xs,
Max 90 -xx + 100 x2 + 80 x 3 ; S.t. 0.92 xx + 0.97 x2+ 1.04 * 3 = 0.98
440 xx + 490 x2 + 480 x2 = 450
13 16
x1 +x2+xs= 1 (i.e., 100%)
100 * 1 + ioo : 100 '
>0
9. Mona Food Company produces four types of foods whose contents are given
below :
Content per unit Cost per unit
Food type in Rs.
Protiens Fats Carbohydrates
Wheat - apple 3 2 6 45
Wheat - banana 4 2 4 40
Wheat - honey 8 7 7 75
Wheat - orange 6 5 4 65
Minimum Requirement 80 40 70

Formulate into LPP.

Answer : Min Z = 45 xx + 40 x2 + 75 x 3 + 65 x4
S.t 3 xy + 4 x2 + 8 x3 + 6 x4 > 80 ; 2xr + 2 x2 + 7 x3 + 5x4 > 40

6 xj + 4 x2 + 7 x3 + 4 #4 > 70 ; x1? #2> X3> #4 ^ 0

46 Operations Research : Theory and Practice

10. Sanja 8c co. manufactures three products A, B and C by using three machines
Milling, Lathe and Grinder. The information related to this manufacture is
given in the following table. Formulate for LP.
Machine Product^ Product B Product C Available hours
per week
Milling machine 8 2 3 250
Lathe 4 3 1 150
Grinder 2 1 1 100
Profit contribution 20 6 8

Answer : Max Z = 20 Xj + 6 x2 + 8 x 3

S.t. x 3 <25 < 150

100; x3 > 0

11. Mother Love Engineering College has to decide its monthly investment plan
for Rs. 10,000 on sports and games. It has three disciplines Mechanical
Engineering (ME), Computer Science Engineering (CSE) and Electronics and
Computer Engineering (ECE). Each discipline has got about 60% boys and 40%
girls. The probability of winning in the inter college sports meet in various
events is as follows.
Discipline ME(boys) ME(girls) CSE(boys) CSE(girls) ECE(boys) ECE(girls)
Probability 0.3 0.25 0.35 0.4 0.5 0.45

It is the policy of the college that at least 40% of funds are to be invested on
ME, and not more than 35% on CSE or ECE branches. Also at least 30% is to
be invested on girls. Formulate the problem into an LPP.
Answer : Let xXB investment on ME boys,
xlG investment on ME girls,
x2B investment CSE boys,
x2G investment CSE girls and so . . . on.
and maximise the total probability of winning
Max Z = 0.3x l B + 0.25x lG + 0.35 x2B + 0.4 x2G + 0.5 xSB
S.t. xlBlB ++ xxlGlG ++ xx2B2B ++ xx2G
2G + xw + x 3G < 10000

X
$B + X
SG - 3 5 0 0 ; 3000
X
\B + X
2B - 7000
X
\B > X\G ' X2B > X2
Formulation of Linear Programming Problems 47

12. A second hand automobile dealer wishes to maximises his profits on his sales
and purchase of cars. He can purchase three types of cars Maruthi 800CC,
Maruthi-Zen and Maruthi-1000CC at Rs. 50,000, Rs. 70,000 and Rs. 80,000
respectively. These can be sold at Rs. 60,000; Rs. 80,000 and Rs. 1,05,000
respectively. The sales probability of the cars in this month are 0.7, 0.8 and 0.6
respectively. For every two cars of Zen, he should buy one car of 800CC or 1000CC
and he has 10 lakhs to invest. How should be invest? Give formulation only.
Answer: xy = investment on 800 CC
x2 = investment on zen and
x$ = investment on 1000 CG,
Profit = 2*, ( sales - profit) sales prob.
Max Z = 10000 xj x 0.7 + 10000 x2 + 0.8 + 250000 x 3 x 0.6
or Max z = 7000 xx + 8000 x2 + 15000 x3
S.t. 70000 (2x2) + 80000 x3 < 1000000 or 14 x2 + 8 x3 < 100
xlfx2, x3>0
13. An Engineering College needs 22 to 30 Professors/Lecturers daily depending
on the teaching loads and lab work distribution^. The requirements are as
follows:
Time Period Number of Professors/Lecturers needed
9 am to 11 am 22
11 am to 1 pm 30
1 pm to 3 pm 25
3 pm to 5 pm 23

The College has employed 24 staff members on permanent rolls and needs
some part time professors. The part time professors must put in exactly 4 hours
per day but can start at any time between 9 am to 1 pm. Permanent employees
(full time lecturers) work for 9 am to 5 pm but are allowed one hour for lunch.
Assume half of full time lecturer take lunch at 12 noon while other half at 1 pm
and thus full timers provide 35 hours per week of teaching time. The university
limits the college to engage part time hours to a maximum of 50% of the day's
total requirement. Part timers earn Rs, 280 per day while full timers earn Rs.
900 in salary and benefits on an average. The management has to set a schedule
so as to minimise the total daily manpower costs.
Answer: Let x± = full time lecturers/professors
x2 = part times starting at 9 am
x3 = PTs adding at 11 am and
= PTs adding at 1 pm
Min Z = 900 280 (x
48 Operations Research : Theory and Practice

S.t xY + x2 > 22 for 9 am to 11 am

— xx +x 2 + *3 > 30 for 11 am to 1 pm

— xj + ^3 + x4 > 25 for 1 pm to 3 pm

acj 1- x4 > 23 for 3 pm to 5 pm

4 (*! + x2 + x3) < 0.50 (22 + 30 + 25 + 23)
(part - timers should not exceed 50% of total hrs).
xx, x2, xs, x4>0

14. Mr. Raju, a retired Govt. officer, has recently received his retirement benefits
viz. PF, gravity etc. He is contemplating as two how much he should invest in
various alternatives open to him so as to maximise ROI. The investment
alternatives and and subjective estimate of risk on five point scale is given. The
relevant data about the investment, return and no. of years blocked is as follows.
Investment Return No. of Years Risk
Govt. Securities 6% 12 1
Company deposits 13% 3 3
Time deposits 10% 5 2
Equity shares 20% 6 5
House construction 25% 10 1

Mr. Raju wants invest to maximise ROI so that the risk is not more than 4 and
funds should not be blocked for more than 15 years. He should necessarily
invest at least 28% in house construction. Formulate the data as an LPP.
[OU - MBA Dec. 2000]

Review Questions
1. Explain the structure of an LPP with examples.
2. Discuss different types of constraints that occur in LPP
3. Discuss the significance of condition of variables in LPP
4. How do you set objective function of an LPP
5. Define linear programming problem. Give example
[0U - MBA Feb 93, M 95, A99]

6. Discuss applications of LPP [JNTU CSE 95, Mech. 97/C, 98/P, 99/S,
0U - MBA - M 95, F 93, A 94, M 92, S 2001, J 2000]
Formulation of Linear Programming Problems 49

Objective Type Questions

1. Which of the following is not a part of an LPP problem
(a) objective function (b) constraint set
(c) condition set (cl) end corrections
2. The availability constraint is represented with
(a) < (b) <
(c) > (d) >

3. A constraint represented by '=' sign is

(a) availability, constraint (b) requirement constraint
(c) exact constraint (d) no constraint

4. If a maximum limit is imposed on a resource usage then it is known as

(a) availability constraint (b) requirement constraint
(c) exact constraint (c) inside constraint
5. A product P contributes a profit of Rs. 50 and product £) contributes Rs. 60.
This is represented in objective functions Z as for xp and x^ products
manufactured.
(a) min. Z = 50 xp + 60 XQ (b) max. Z = 60 xP + 50 XQ
(c) min. Z = 60 xp + 50 XQ (d) max. Z = 50 xp + 60 XQ

6. The products requires 2 man hours, B requires 3 hrs. and C requires 4 man
hrs. If the company works for 6 day week with only one shift of 8 hrs, then this
is represented as (Assume x^,x2, x3 units are manufactured in
A, B and C types respectively)
(a) 2xx + 3x2 + 4x3 < 56 (b) 2xx + 3x2 + 4x$ > 8/6
(c) 2x} + 3x2 + 4x3 > 48 (d) 2xx + 3x2 + 4x3 < 48

7. A company has to decide how many units of each product to be produced. It

estimates Rs. 60perunitofproductXandRs. 80 per unit of product Y. It incurs
cost of Rs. 40 on X and Rs. 30 on Y to manufacture. The objective function can
be written as
(a) max. Z = 60 x + 80^ (b) max. Z = 20 x + 50y
(c) max. Z - 30 x + 40 y (d) min. Z = 40 x + 30 y
8. If a product Xj contributes a profit of Rs. 75 per unit and product X2 costs Rs.
50 per unit, the objective function/(x^) is written as
(a) min./(x) = - 75 xx - 50 x2 (b) m a x . / M = 75xL + 50 x2
(c) max./(x) = 75 xx - 50 x2 (d) max./(x) = 25 a-
50 Operations Research : Theory and Practice

9. The non-negative conditions for variables xx and x2 can be represented as

(a) JCJ > 0, x2 > 0 (b) xi = x 2 ^ 0

(c) xj > 0, x2 > 0 (d) Any of the above

10. A furniture manufacturer produces x{ chairs and x2 tables every week. He has
an order for 50 chairs and 60 tables in a week. This is represented as
(a)x 1 +x 2 > 110 (b)max. 5 0 x ^ 6 0 x2
(c) *x < 50, x2 < 60 (d) xx > 50, x2 > 60

11. If your pocket money is Rs. 100/- per day, assuming Rs. x you spend on a day,
formulate it.
(a) max. 100 x (b) min. 100 x
(c)x>100 (d)x<100

12. You are purchasing glass for spectacles. The variable is the power of the glass
to suit eye sight and size of the glass. The constraint used in this regard is

(a) less than or equal to (b) greater than or equal to

(c) exact type (d) can not be represented
13. Production of x pens cost Rs. 100 and that of y pencils cost Rs. 120. If the
company wishes to optimise unit costs, the objective function is

(a) min. 100 x + 120 3> (b) min.

(c)max.- — - — (d)x< 100, y < 120

J
x y
14. 10 units of product xj costs Rs. 40 while 5 units of product x2 costs Rs. 25. The
objective function of this information will be
(a) min. - 10 xY - 5 x2 (b) min. 40 xx + 25 x2
(c) min. - 4xj - 5x2 (d) min. 4xj + 5x2

15. A belts of type -1 requires leather twice as that of type - II. Total leather available
is sufficient for 300 belts of type II. If xx and x2 are the number of belts of type
I 8c II to be produced respectively, then the constraint is
(a) 100 xx + 200 x2 < 300 (b) 200 xx + 100 x2 < 300
(c) xi = 2x2 < 300 (d) none of the above
16. Which of the following is usually minimised
(a) sales (b) profits
(c) operating time (d) revenue
Formulation of Linear Programming Problems 51

17. The requirement constraint is shown with a mathematical relationship as

in LPP formulation
c
(a) axy + bx^ - (b) ax\ + bx9 > c
(c) axy + bx% ~c (d) c > axy + bx%

18. Cost of 5 pens is equal to the cost of 8 pencils. I have to maximise the number
of pens and pencils (say Xy andx 2 respectively) when I have Rs. 100/-. The
problem is formulated as
(a) max. x1 + x2 subject to bxy + 8x2 < 100
(b) max. xy + x2 subject to 5xy + 8x2 > 100
(c) max. 5xj + 8x2 subject to Xj + x2 = 100
x} x2
(d) max. xx + x2 subject to -— + —-< 100
D o

19. A machine is to be stopped for at least 30 min, if it continuously runs for 2

hours. If x is the number of machine hours to be planned on this machine in a
8 hour day, we have constraint as
(a) x < 8 (b) x < 6
(c) x < 6.5 (d) x > 6.5

20. A variable unrestricted in sign can assume the values as

(a) only positive (b) only negative
(c) zero (d) any value.

Fill in the Blanks

1. The objective function in an LPP will be in the form of or
2. The objective function chosen for an LPP when costs are given is
type.
3. A constraint, mathematically represented as ax + by = c is called
type constraint
4. If a minimum limit is imposed on production of a certain commodity, the
constraint is said to be type
5. The cost a product X is Rs. 5 per unit while that of product Y is Rs. 8 per unit.
If x and)? are the units to be manufactured, the objective function is

6. Krishna has Rs. 100 in his purse and went to a restaurant. He orders Xy number
of units of worth Rs. 10 each andx 2 number of units of worth Rs. 20 each. Then
constraint is written as
52 Operations Research : Theory and Practice

7. The number of units of salt to put in a curry dish, to maximise its taste is an
example for type of constraint

8. A company has to spend Rs. 200 per unit of advertisement in New paper while
Rs. 400 per unit in T.V. If the budget is Rs. 10,000 then the constraint is written
as

9. Product M costs Rs. 10 and product TV cost Rs. 40 per unit manufactured. The
profit contribution are Rs. 50 and 30 respectively. The objective function is

10. Products P and Q cost Rs. 10 and 20 each to produce. The objective function
is given as max. Z = - 10 xy - 20x 9 wherex^ =

Answers
Objective Type Questions :

l-(d) 2.(b) 3. (c) 4. (a) Md)

6.(d) 7.(b) 8. (c) 9.(c) 10. (d)

11. (d) 12. (c) 13. (c) 14. (d) 15. (d)
16. (c) 17. (b) 18. (d) 19. (c) 20. (d)
Fill in the Blanks :
1. Maximisation, Minimisation 2. Minimisation
3. Exact or equality 4. Requirement

5. min. Z = 5x\ + 8x£ 6. lOxi + 20x2 < 100

7. Availability or < 8. 200xi + 400x9 < 10000

9. max. z = 40 x\ - 10x2 10. No. of products of P to be produced,

No. of products of Q to be produced
Graphical Solution of LPP 55

3.0 Introduction
In the previous chapter you have learnt how to formulate a given information
into a linear programming problem. In this chapter you will learn how to solve these
problems using graphs.

3.1 Merits and Demerits of Graphical Solutions

Graphical solutions are easier to understand and reproduce. Also a pictorial
view is always a better representation. Thus graphical solutions have gained
prominence in Operations Research.

However, graphical solutions have certain limitations such as :

1. Limited to the problems of two decision variables only.

2. Accuracy can not be obtained.

3. Some times it is difficult to represent certain expressions, particularly in the

case of non linear expressions.

3.2 Graphical Solution Procedure

1. Assuming xj and x2 (the decision variables) to be represented on X and Y axes
respectively check the conditions of variables and accordingly prepare the
graph sheet.
If xj > 0, x2 > 0, your graph will be in first quadrant, if xj is unrestricted and
x2 > 0, the solution will be I and II quadrants, if Xj > 0 and x2 is unrestricted
the solution will be in I and IV quadrants. If both are unrestricted, the solution
may be any where (in any quadrant).
[It is better to leave three units at the bottom and also at the left side so that your graphical
solution will be clearly represented].
Divide the scale approximately on X and Faxes such that you can represent all
its values. It is always better to have same scale on both axes.
2. Assume the constraints as equations and find any two points for each equation
so that the equation can be represented as a straight line on graph.
[It is always easy to assume x2 =0 to find x^ (say a) and Xy = 0 to find x2 (say b) and
thus you draw line connecting (a,0) and (0,b)]
3. Similarly, draw all the constraint lines.
4. Shade the appropriate areas as given by the constraints. If the constraint is <
type shade the area towards origin. If the constraint is > type shade the area
away from origin. If the constraint is ' = ' type, do not shade any area, and the
line itself is the region.
56 Operations Research : Theory and Practice

Note : The above rides are applicable for the lines which have positive constant at right hand side, otherwise
the above rides are to be reversed. Refer section 3.5 for details.

5. Identify the feasible region by locating the area satisfying all constraints that is
common with all the constraint areas. This region will have the optimal
solutions. If there is no common area possible then the solution is infeasible.

6. Now, to find optimal solution we have two methods :

(a) Iso-profit (or Iso-cost) line method.

(b) Extreme point evaluation method.

(a) Iso profit line method for maximisation objective [Iso cost line in the case
of minimisation objective]: In this method, first assuming zero profits or
zero costs, the objective function line is drawn. (Find any one point giving
any value to x± and find x2. Connect this point to origin and extend). Any
line parallel to this line belongs to the family of objective functions. Among
all these family of lines locate the one that will be touching the feasible
region at the farthest point (in the case of maximisation) or the one which
touches at nearest points (in the case of minimisation).Identify these
points and read their X and Y co-ordinates.

Find the Z m a x or Z m i n (as the case may be) by substituting these values in
the objective function.

(b) Extreme Point Evaluation Method : In this method, after locating the
feasible region, read all the extreme (corner) points. Substitute these values
in objective function. Identify the point(s) that yields maximum value or
minimum value (as the case may be). Hence the solution.

[This method is easier (comparatively) but not reliable in some cases. Therefore
students are advised to use the Iso-profit/cost line method and may check the answer
by using extreme point evaluation]

These are illustrated through the solved examples to follow.

Illustrations of Graphical Method :

ILLUSTRATION 1 — — ^™________™_— „„_„

Solve graphically
Maximise Z=8x1 + 6x2
Subject to 2xj + x2<72
Xj + 2x2 < 48
X!>0,X2>0 f JNTU B.Tech Mech 93]
Graphical Solution of LPP 57

Step 1 The graph sheet is prepared selecting x} on X - axis, and x2 on Y - axis. It is

clear from the conditions (x1 > 0, x2 > 0) that we get the figure in first quadrant.
Step 2 df 3 Write inequations as equations and locate any two points on each equation.
2x{ + x2 < 72 written as 2xx + x2 = 72
If xy = 0, x2 = 72 => point is A (0, 72)
and xy = 36 if x2 = 0 => point is B (36, 0)
Again Xj + 2x9 < 48 is written as x^ + 2x9 - 48
If xj = 0, x2 = 24 => point is C (0, 24)
and xj - 48 if x2 = 0 => point is D (48, 0)
Draw these lines on the graph sheet as AB with the points A (0, 72), B (36, 0)
and CD with points C (0, 24) and D (48, 0)
Step 4 Shade the corresponding areas that represents the inequations, i.e. shade the
area below AB towards origin since the inequation is < type, similarly below the
line CD
Step 5 Locate the common area that can satisfy all the constraints and name it as
"feasible region". In the following graph it represented as OCEB
Scale
OnXaxis-(xl)
lcm = 10 units
on Y axis (x2)
lcm = 10 units
Legend
, x 2 <72
Parallel obj. x, + 2x2 < 48
func. line through
max. point. E(32, Feasible region
(AECO)
x2 - 72 (AB)

\
x
l 0 36
F(-30, 4 0 ) \
X
2 72 0
\ x, + 2x2 = 48 (CD)
x
l 0 48
x2 24 0

FIGURE 3.1 :
58 Operations Research : Theory and Practice

Step 6 : (a) Iso-Profit Line Method : The objective function line is first drawn through
origin treating profit Z - 0.
Z o = 8x{ + 6x2 = 0 passes through 0(0, 0)
Since if xx = 0, x2 - 0
Also, if x2 = 40 => Xj = - 30
F (-30, 40) lies on objective function line.
OF represents Sxy + 6x2 = 0.
Represent this line with'dot and dash' to distinguish from the constraint lines.
Now draw a parallel line by using set - squares so as to touch at the most
maximum point on feasible region. This is E (32, 8).
Thus Z m a x = 8 (32) + 6(8) = 256 + 48
= 304 units.
(b) Extreme Point Evaluation :
The evaluation is shown in a tabular form given below :
Extreme x
i x2 8^ fcc2 Z = Sxl + 6x2
Point
0 0 0 0 0 0
C 0 24 0 144 144
E 32 8 256 48 304 <r- Z m a x
B 36 0 288 0 288
Solution
=

ILLUSTRATION 2
Find the minimum value ofZ= 4xt + 2x2 subject to the constraints:
x1+2x2>2, 3xj+x2>3, 4xj+3x2>6and
xD x2>0 by graphical method. [JNTU Mech./Prod.Chem. 2000]

Solution i
First convert the inequality into equality and find the minimum two
co-ordinates to draw a straight line in the graph as follows:
Constraint 1 : x± + 2x2 = 0 Constraint 2 : + x2 = 3

0 2 0 1
1 0 x2 3 0

Constraint 3 : 4xY + 3x2 = 6 Objective function line 4xx + 2x% = 0

0 —3 x
i 0 -1
x2 2 6 H 0 2

Assume Z = 0 «'.«., 4xj + 2^ 2 = 0 to

^nc? two co-ordinates for this equation.
Graphical Solution of LPP 59

Now the geometrical interpretation of the problem is given in graph. (Fig. 3.2)

Feasible
region

\ B(2,0)
F(1.5,0)

FIGURE 3.2 :

The minimum value of Z is found by a parallel lines to Z = 0. The minimum

point which it touches the feasible region is solution i.e., at P (0.55, 1.25).

Min Z = 4XX + 2X2 = 4 x 0.55 + 2 x 1.25 = 4.7

Note : Observe the shadings given in illustration 1 and 2, In illustration 1 the shading is done below the
constraint lines since they are of< type, while in illustration 2 the constraint are > type, therefore
the areas above these lines are shaded. Similarly in the illustration 1 objective function is
maximisation type therefore we take highest point on the feasible region that objective function can
touch, whereas in illustration 2 we take the lowest point touched by objective function on the feasible
region since it is to minimise.

ILLUSTRATION 3

An oil refinery can blend two grades of crude oil to produce quality P and Q
petrol. Two possible blending processes are available. For each production run,
the olderprocess uses 5 units of crude A, 7 units of crude B and 2 units of crude C
to produce 9 units ofP and 7 units ofQ. The newer process uses 3 units of crude
A, 9 units ofB and 4 units of crude C to produce 5 units ofP and 9 units ofQ
petrol. Because of prior contract commitments the refinery mustproduce at least
500 units ofPandat least 300 units ofQ for the next month. It has available 1500
units of crude A, 1900 units of crude B and 1000 units of crude C. For each unit
ofP the refinery receives Rs. 60 while for each unit ofQ it receives Rs. 90. Find
out the linearprogramming formulation so as to maximize the revenue and solve
graphically or other wise.
Solution :
Refer Illustration 7 of chapter - 2 for formulation of LPP,
Summary : The above information is formulated as given below :
Maximise Z= 1170x2 + 1110x2
60 Operations Research : Theory and Practice

Subject to 9xl + 5x2 > 500 500 -i

1x} + 9x 2 > 300

5 x , + 3 x 9 = 1500
5x1 + 3x2 < 1500
lxx + 9x2 < 1900
2x} + 4x2 < 1000 9x, + 5x2 = 500
x^ t x% > 0

The feasible region is formed by

two constraints only. These are
9xx + 5x2 > 500 and 7xx + 9x2 < 1900.
The other three constraints are
redundant constraints.
Solution by Extreme Points :
z =1170 3:!
Point *1 x2 1170 X x 1110X 2
+ inox2
A 0 211.11 0 234332.1 234332.1
B 0 100 0 111000 111000
C 55.55 0 64993.5 0 64993.5
D 271.43 0 317573.1 0 317573.1 ^ 2 m a x

Solution .
X x = 271.43 Z
X2= 0 max = 317573.1

3.3 Redundant Constraint

It is the constraint without which there will not be any effect on the feasible region.
Ex: 1. Refer solution of above p r o b l e m . T h r e e constraints given by
7xY + 9%2 > 300, 2x} + 4x2 < 1000 and 5x} + 3x2 < 1^00 are redundant.
2. Suppose you have a constraint x^ + x2 > 2 and another constraint x± 4- x9 > 10,
then the first one is redundant since second constraint if gets satisfied,
automatically satisfies the first.
3. If 2#i + 3#2 - 5 0 and 4x^ + 6x2 < 37, the former is redundant.
ILLUSTRATION 4
Use graphical method to solve thefollowing LPproblem:
Maximize Z=3Xj+ 4x2
Subject to the constraints;
2x2+x2<40
2x1 + 5x2<180
XJf x2>0 [JNTU B. Tech (Mech) 1998/S, OU - MBA July 2000]
Graphical Solution of LPP 61

Solution ;
Consider a set of rectangular cartesian axes 0Xx and 0X 2 in 0XY X2 plane. Plot
the lines by considering the inequations as equations.
2xx +x 2 = 40 is ploted by taking two points A (20, 0) and D (0, 40) and
2xt + 5*! = 180 is ploted by taking the points B (0, 36) and E (90 , 0).
Now, shade the region to represent the inequations (below the line for < ) and
to find the solutions space or feasible region. The feasible region is OACB as shown
in graph.
The four vertices (corner pionts) form a convex region. These are O(0, 0),
4(20,0), B (0, 36), C (2.5, 35).

Now, on substituting these D(0,40)_ -2 X l +x 2 = 40 2x,+x 2 <40

B
points in the objective function (°> 36)\ 2.5,35) 2x, + 5x2<180
Feasible region
z0 = 0 + 0 - 0
zA = 3 x 20 + 0 = 60
zB = 0 + 4 x 36 = 60 2x, +5x 2 = 180
zc = 3x2.5 + 4 x 3 5 = 147.5
zc is maximum,
Hence, the optimum solution is
xx = 2.5
x2 = 35 1 I
zmw - 147.5 O(0, 0) A(20, 0) E(90, 0)
FIGURE 3,4 :
ILLUSTRATION 5
Find the minimum value ofZ = — x1 + 2x2
Subject to the constraints
-Xj+3x2<10
Xj+x2<6
xr-x2<2
x19 x2>0
Use graphical method [JNTU B.Tech (ECE) 1997]

Rewriting each constraint as follows : (in the equation form)

-xY + 3x2 = 10

X\t #2 ^ 0
Draw line for each constraint by first treating it as a linear equation. Then use
the inequality condition of each constraint to make the feasible regions as shown in
the figure.
62 Operations Research : Theory and Practice

FIGURE 3.5 :

The coordinates of extreme points of the feasible region are

O = (0, 0)
A = (2, 0)
B = (4, 2)
C = (2, 3.8)

The value of the objective function at each of these extreme points is as follows:

Coordinates Objective Function Value

Extreme Point
(*1 9 X2 )

O (0,0) -1 (0) + 2(0) = 0

A (2,0) -1 (2) + 2(0) = -2
B (4,2) -1 (4) + 2(2) = 0
C (2, 3.8) -1(2) + 2(3.8) = 5.6
D
(••fl
The minimum value of the objective function Z = - 2 occurs at the extreme
point (2,0). Hence the optimal solution to the given LP problem is Xj = 2 , x% = 0
and Min Z = - 2.
Graphical Solution of LPP 63

3*4 Different Types of Solutions

As we have seen the structure of an LPP is composed of three main parts viz
objective function, set of constraint and the conditions of variables. With reference to
these parts, we report the following types of solutions.
1. Solutions.
2. Feasible solutions.
3. Basic feasible solutions.
4. Optimal solutions.
1. Solutions : All those values of variable which satisfy the conditions given are solutions;
Thus if the conditions are Xy>0, x9 > 0 (non - negative), then all the values in
first quadrant (covered by positiveX- axis and positive Y- axis) are the solutions.
Similarly, x 1 > 0 (X-axis) and x 2 unrestricted yields the solutions in first and
fourth quadrants of graph, and so on.
2. Feasible Solution : All the solutions which satisfy all the conditions of variables as
well as all the constraints are feasible solutions.
We can notice here that if all the constraints are exact type, we may get 'point solution'.
One exact and another inequality constraint will yield 'line of solutions'. All the inequality
constraints will generate an area of feasible solutions often referred to as feasible region.

a2x + b2y < c9

+ b 1 y = c1
Line of solution
Point solution
a 2 x 2 + b 2 y = c2 a 1 x+b 1 y =

(a) Point solution with two exact constraints (b) A line of solution with one exact constraint
(point size feasible region) (line feasible region)

a 2 x+b 2 y<c 2

y < c,

V\
(c) A feasible region with inequality constraints
FIGURE 3:6 :
64 Operations Research : Theory and Practice

3. Basic Feasible Solutions : The values of variables represented by the points along
the border lines offeasible region are basic feasible solutions.
If m non identical equation with V variable (m < n) exist in a problem, then
keeping (n - m) variables constant, usually at zero, the values of the variables yield a
solution called lbasic solution1. As it satisfies all the constraints it can be called a region,
it can also be called a 'basic feasible solution'. Therefore each selection of (n - m)
variables from n variables gives raise to l(nr ^' basic feasible solutions' (BFS)
I "-")
Of all these BFS, the one with which we start working out the problem is 'initial
basic feasible solution'. (IBFS).
Most commonly, the IBFS will be choosen to start at the worst case of the
solution set so as not to miss to examine any solution. Thus a solution with zero profit
or nil production or the values of decision variables as zeros (i.e., origin) in graphical
solution will be IBFS.

4. Optimal Solution : The solutions which satisfy all the conditions of variables,
all the constraints and the objective function are 'optimal basic feasible solutions
(OBFS)' or simply 'optimal solutions'.
Schematically:

Solution - Conditions of variables

Feasible solution - -*• Constraint set

Optimal solution - 5 Objective function ^H

FIGURE 3.7 : TYPES OF SOLUTIONS

3,5 Locating Feasible Region

Most of the students of Operations Research are complaining that they are
getting conflised in locating a feasible region. Some tips are given below to remove
their confusion.

. ax,+bx, >C x 9 >C

ax, + bx2 < C

\
(b) Imagine the line of fig. (a) is
(a) Standard form rotated at point 'B'
Graphical Solution of LPP 65

t - bx 2 > C

j - bx 2 < C

0
(c) Imagine line in fig. (a) is further rotated at 'B' (d) Imagine line of fig. (a) is rotated at 'A'
-ax1 + bx2 < C
(of)axl - b x 2 .
Note:
- bx2 > C
t
In a simple way youcan rember as
'or) axj - bx2 < C
1. If area is comming towards origin, it is <
type constraint, and
2, If area is going awayfrom origin, it is >
type constraint provided the right hand
side ofthe inequation is with positive sign.

(e) Imagine line in fig. (a) is further rotated at point 'A'

FIGURE 3.8

3.6 Types of Feasible Regions

The area consisting of set of feasible solutions exist in several ways. These are
described below :
1. With Reference to Feasibility :
(a) Infeasible Region : If there exists no feasible region (solution) at all, it is
infeasible region.
If there, is no common area or line or point that satisfies all the constraints,
the solution is infeasible as no feasible region exists,
(b) Feasible Region : It is a point or line or region that satisfies the set of
constraints.
2. With Reference to Number of Feasible Points :
(a) Infeasible Region : As discussed above.
(b) Point Feasible Region : If only one point exits as feasible region e.g., Two
equality constraints intersecting at a point and satisfying the conditions of
variables (or three or more equality constraints concurrent at a point).
(c) Line Feasible Region : It is a feasible region in the form of a line. It is found
in case at least one equality constraint exist in the set.
66 Operations Research : Theory and Practice

(d) Area Feasible Region : It is the area surrounded by many inequality

constraints.
(e) Volume Feasible Region : It is the feasible region found when there are three
variables by which three planes are formed. It is three dimensional.
3. With Reference to Shape :
(a) Convex Feasible Region : It is a region formed by a collection ofpoints such
that every point on a line segment dram selecting any two points of the
collection, falls in the collection. This means that if P and Q are any two
points in a collection, the line segment PQ is also in the collection.

FIGURE 3 . 9 : CONVEX FEASIBLE REGION

(b) Concave Feasible Region : It is a region formed by a collection of points
such that at least one point on a line segment drawn selecting any two points
of the collection, does not fall in the collection. This means that if
P and Q are any two points is a collection, at least one point on line segment
PQ does not fall in the collection/region.

FIGURE 3 . 1 0 : CONCAVE FEASIBLE REGION

4. With Reference to its Boundaries :

(a) Bounded Feasible Region : This feasible region will have a distinct
boundary line. Most times the boundaries are connected by the co-ordinate
axes and some availability (< type) constraints.

Bounded feasible region by

abscissa, ordinate and
two availability constraints

FIGURE3.il : BOUNDED FEASIBLE REGION

Graphical Solution of LPP 67

(b) Island Feasible Region : This is also a bounded feasible region but not
connected by the co-ordinate axes.

FIGURE 3 . 1 2 : ISLAND FEASIBLE REGION

(c) Unbounded Feasible Region : This feasible region contains no boundary on

atleast one side. Usually it is observed with all requirement, (i.e., > type)
constraints.

Unbounded
feasible region

FIGURE 3 , 1 3 : UNBOUNDED FEASIBLE REGION

3,7 Types of Optimal Solutions

The optimal solutions can be categorised into the following types:
1. Infeasible solutions.
2. Unique solutions.
3. Multiple optimal solutions,
4. Unbounded solutions.
1. Infeasible Solution : When feasible region does not exist, the solution we get is
infeasible.
In graphical solution it is found when one constraint is availability (<)• type and
the other is requirement (>) type and these two can not produce any common area
(non intersecting) in the specific quadrant (such as x± > 0, x%, > 0 indicates first
quadrant).
68 Operations Research : Theory and Practice

Requirement constraint
(a2x + b2y > C2)
Availability constraint

FIGURE 3 . 1 4 . : INFEASIBtESOLUTION

2. Unique Solution : If the optimal solution is one and only one, then the
solution is unique.
Thus the feasible region will yield one maximum or minimum solution valvies
in the objective function at only one point.

\
\ Unique (maximum) Unique (minimum)
solution point solution point
\
\

(a) Maximisation (Unique) solution (b) Unique solution with minimisation

F I G U R E 3 - 1 5 : UNIQUE SOLUTIONS

3. Multiple Optimal Solution : When several optimal solutions exist, the

solutions are said to be multiple optimal solutions or alternate optimal
solutions.
In graphical solutions they exists when the objective function is parallel (having
same slope) to one of the constrains, provided there exists a feasible region.
a2x + b 2 y < C 2 a2x + b 2 y > C 2
Any point along this portion
. of line is optimal Any point along this
portion of line is optimal

'#//„.

(a) Multiple optional solution (maximisation) (b) Multiple optional solution (minimisation)
FIGLiRF. 3 . 1 6 : MULTIPLE OPTIMAL SOLUTIONS OR ALTERNATE OPTIMAL SOLUTIONS
Graphical Solution of LPP 69
4. Unbounded Solution : If a distinct and finite solution can not be found or
the solution exists at infinity, the solution is said to be unbounded.
In graphical solution, unbounded solutions are obtained if the feasible region
is unbounded (formed by requirement constraint i.e., > type) while the objective
function is maximisation.

Unbounded solution

FIGURE 3 . 1 7 : UNBOUNDED SOLUTION

(Since it has to be taken to infinity to locate maximum value we have no finite
or unbounded solution)

Solved Problems
Infeasible Solutions ———
ILLUSTRATION 6
Solve Graphically:
Maximise
Subject to Xj+x2<12
2x1 + 3x2>60
[ JNTU B.Tech (Mech) 96/CCC]

Solution : = 12 2xl = 60

0 12 0 30
x2 12 0 x2 20 0

Scale
On x axis
1 cm = 5 units
on y axis
1 com = 5 units

FIGURE 3 . 1 8 :
70 Operations Research : Theory and Practice

Unique Solutions <

In addition to the above illustrations (1 to 5), one more is given below.
ILLUSTRATION 7 — — ——— — —
Minimize Z = 2x}+ 3x2
Subject to X;+x2<4, 3Xj+x2>4
Xj + 5x2>4, 0<x1<3, 0<x2<3
(Remark: In this problem the conditions ofvariables are given in constraints).

Points for xx + x2 = 4 are (0, 4), (4, 0)

[4
Points for 3xY + x2 = 4 are (0, 4), — 0
f 4V
Points for X| + 5x2 = 4 are 0, — L (4, 0)
and x} - 3 is a line parallel to X2 - axis i.e., Y axis passing through (3, 0)
Similarly x2 = 3 is a line parallel to X1 - axis i.e., X - axis passing through (0, 3)
Also, 0 <xj or xx > 0 and 0 <x 2 or x2 > 0 represent the condition that the
solution exists in first quadrant.
Now these can be plotted on the graph as shown below :

(-3,2)

-2 + C(|,0)

FIGURE 3. i 9 :
Feasible region is area covered by GHIJK
Iso - Cost Line Method : Bringing Z = 2* ^ + 3#2 t o tne
minimum point of
feasible region it touches at G.
( 8 -i
Therefore solution is G — —
Graphical Solution of LPP 71

Extreme Point Evaluation :

1
At

ZH=2x3+3x 1=9
Z7 = 2 x 1 + 3 x 3 = 1 1

1 + 3 x 3 = 9.67
is Z min , i.e., 4
u 1 4
At x^ = — a n d x 2 = ~z

Multiple Optimal Solutions — — ^ — ^

ILLUSTRATIOry 8
Maximise 15xj + 9x2
+ 2x2 < 10 , ^0

Solution :

) - , A(0, 10)

10

FIGURE i:i0 :
72 Operations Research : Theory and Practice

As Zo = 15xj + 9x2 is parallel to 10xj + 6x% = 60, it fits exactly on this line when
its iso-profit line is taken to represent maximum. Therefore any point along the line
segment BE is a solution.
Hence there are Multiple optimal solutions to this problems
Extreme Point Evaluation :
At O (0, 0)

Z o = 1 5 ( 0 ) + 9(0) = 0

At C (0,5)

Zc= 15(0)+ 9 (5) = 45

At B (6, 0)
ZB = 15 x 6 + 9 x 0 = 90 (Maximum)

30 20

= —— = 90 (Maximum)

As the values at B and £ are showing maximum, we can infer that any point
along the line BE is optimal and hence the solution is multiple optimal.
Remarks:

(i) You may be misled if you misread the values ofE in the graph.

Suppose you read as

X] = 4.3 andxg = 2.9
ZE = 15 (4.3) + 9(2.0) = 64.5 + 26.1
= 90.6-91
Thus it may give a wrong anser as 'unique solution' at E with Z m a x = 90.6 or 91.
Therefore it is very iviportant to read exact values (which is very difficult in graphical solutions).
However, the exact values can be obtained by solving the equations tofind,the point ofintersection.
(ii) In view of the above, Students are advised to choose the Iso-profit/Iso-cost line method to
eliminate such error.

[ 15 —5 ^
- — = — and that oj first constraint
y 5
i.e., = - —• I As these are same by looking at the problem, we can predict the multiple
v . J
optimal solutions.
Graphical Solution of LPP 73

(iv) The above remark fails if there is no chance of getting feasible region. (Infeasible solution)
(v) The remark (iii) also fails if the objective function is to minimise in this case,
Since Zm[n — 0 at origin is most minimum.
Unbounded Solution — — — — — ^ — ^ ^ ^ - ^ ^ — — —
ILLUSTRATION 9
Solve:
Maximum Z=3x1+4x2
Subject to x}-x2<l
-Xj+x2<2
AJ7+JC2>0 [JNTU(CSE) 3/4 -2001]

.Solution -X2= 1

0 1 x
l 0 _2
-1 0 2 0

\
\

Scale
on X axis
lcm = 1 unit
on Y axis
lcm= 1 unit
Legend
\ o Xj + X2 < 1
\(4,-3)
\ -X! + X2 <
\
FIGURE 3.21 •
74 Operations Research : Theory and Practice

Since there is no boundary on one side and the maximum value of

Z = 3xy + 4x2 is found at this side only, the solution is found at infinity. Thus there is
no finite solution to this problem. In other words, the solution is unbounded.
Remarks :
(i) In the above problevi, if the both equations are greater than or equal to (>) type i.e.,
xx~ x2> 1 and - x± + x2 > 2 then the solution will be infeasible.

(it) Pupils may get confused between unbounded (or no finite) solution and, the infeasible
solution (i.e., no solution). In unbounded solution, the solution exists at infinity, this
indicates infinite scope of improvement (without any finite value) through its objective
function, while in infeasible solution there exits 'no' solution on in other words, it is
impossible to have any solution to the problem.

Practice Problems
Infeasible Solutions — — — — ^ — ^ — — — — —
Solve Graphically :
1. Maximise Z = 8xj + 10x2

Subject to 2xx + 3x2 > 12

xltx2> 0

2. Maximise Z= K k ] + 15x2

Subject to 4xx + 6x2 > 240

xl+x2<A0

4x{ + 3x2 > 90

tfj, X2 > 0

3. Minimise Z = 3xj + 4*:2

Subject to xl ~x2 > 1
~xl +x2> 2
Xj, x 2 > 0

4. Minimise Z = x^ + 2x 2

Subject to xj - JC2 > 3

- Xj + x% > 4

X!>0, x 2 > 0
Graphical Solution of LPP 75

5. Solve graphically, the following LPP if possible.

xx-3x2<§; x1 + 2x 2 >4
x y - 3^2 ^ - 6 and x^ x 2 > 0
With an objective function as
(a) Minimise 2x1+x9
(b) Maximise 2xY + 3x2

Answer : (a) 4 (b) Unbounded.

Unique Solution ^ ^ — ^ — — ^ — ^ — ^ —
Solve Graphically :
1. Maximise Z = 9xI + 16x 2
Subject to xj + 4x 2 < 8 0
2xj + 3x,, < 9 0
x{ > 0, x 2 > 0

Answer : Z m a x = 440 at Xi = 24,; x 2 = 1 4

2. Minimise Z = 3xj + 2x2

Subject to 5>
:j + x 2 > 10
y^ + x 2 > 6
>q + 4x2 :> 12
>^ , x 2 > 0
Answer : Z •„ = 13 atxi, = L x 9 = 5.

3. Solve graphically the following LPP.

Maximise : Z = 3xj + 2x2
subject to the constraints

JCX +x2 < 3

xj < 2 a n d

xj , ^ 2 - ° [JNTU - Mech./Prod./Chem. - 2001/S]

4. Aditi wishes to invest Rs. 12000 in Bank Saving Certificates (BSC) and National
Saving Bonds (NSB). According to rules at least Rs 1000 and at least 2000
should be invested on B.S.C and N.S Bonds respectively. If the rate of interest
on B.S.C is 8% p.a. and that on NSB is 10% p.a., how should Aditi invest her
money to earn highest income. Solve graphically or otherwise.

Answer : BSC Rs. 1000 on NSB Rs 11000

Max. interest = 1180p.a. total income = Rs. 3180
76 Operations Research : Theory and Practice

5. Beetel manufacturers produce two types of telephone sets, viz with cord and
cordless. The cord type requires 2 hours and cordless requires four hours to
assemble. The company has at the most 800 hours per day to assemble and its
packaging department at the most can pack 300 sets per day. If the company
sells cord type at Rs. 600 and the cordless at Rs. 1000, how may sets of each
type should be produced to attain highest sales. Solve graphically.
Answer : cord type = 200, cordless = 100; Max sales = Rs. 2,20,000

6. With the constraints

xl - x9 < 1, xy + x2 > 3, xj > 0, x2 > 0 and the objective as
(a) Minimise Z = 3x| + 2x2
(b) Maximise Z = 3 ^ + 3xr,
(c) Minimise Z = 3xj + 3x2,
Solve the LPP graphically
Answer : (a) x± = 2,x= 1; Min Z = 8

(b) Unbounded solution.

(c) Multiple optimal solution (along the line xj + x 2 = 3)

Xj = 2, x 2 = 1 Z m i n = 9
or xy = 1, x 2 = 2 Z m i n = 9 or x^ = 1.5, x 2 = 1.5 and so on
Multiple Optimal Solution ^ ^ — ^ — « — i ^ — » • — — i
1. Solve graphically. Is the feasible region convex or concave?
Maximise Z - 4xj + 3x2
Subject to 8x1 + 6x 2 <48
2xj<9
x2<6
xj, x2 > 6
Answer : Multiple optimal solutions
x} - 4.5, x2 - 2, Z m a x = 24; or xY = 1.5, x2 - 6, Z m a x - 24
The feasible region is convex.
2. (a) Use graphical method to solve the following LPP.
(b) Examine whether the LPP has any alternative solution, other than you
found.
(c) Also check whether feasible regions is convex/concave and
bounded/unbounded.
Graphical Solution of LPP 77

Minimise Z - 2 0 x ! + 30x 2
Subject to 5
2x 2 > 7
4x1 + 6x 2 < 24
xi >0, x9 > 0

(d) What will be the solution if the objective function in the above problem is
to maximise instead of minimisation.

Answer: (a) x{ = 5, x2 = 2/3, Z m i n = 120

(b) The solution is multiple optimal solutions an alternative solution can

be x} = 0.75 , x2 = 3.5 , Z m i n = 120

(c) The feasible region is concave type.

(d) If the objective function is Z m a x = 20 x] + 30 x2 then the solution is
unbounded i.e., no finite solution.

3. Use the following constraints to solve given objective functions of an LPP.

xj + x9 < 1, x1 + x2 > 3;_ x1 > 0, x2 - 0 a n d objective functions as

(a) Min z = 3 x^ + 2 x2
(b)
(c)
(d) 1 2

: (a) Unique solution x± - 0, x2 = 3, Z m i n = 6

(b) Multiple optimal solutions x± = 0, z2 = 3, Z m i n = 9;

x 1 = 2, x 2 = 1 Z m i n = 9;
(c) Unbounded solution

4. Solve Z m a x = 0.4 xx + 0.2 x2

S.t 2x}+x2<72
xj + 2x2 < 48
xpx2>0

Answer: Multiple optimal solutions Xj = 32, x2 = 8 or x± = 36, x2 = 0

78 Operations Research : Theory and Practice

5. Use graphical methods to solve the following LPP. What type of feasible region
do you get ?
Max Z = 5x + 4y
Subject to x + y < 40
x<20
15x + 12y = 300
x> 0 ,y> 0
;lnm>er: Multiple optimal solution any point along the line 15 x + 12^ = 300
Z m a x = 1 0 0 a t x = 2 0 , y = 0 o r x = 0,y = 2 0 , x = 1 0 , 3 ' = 1 2 . 5 , a n d s o o n .
The feasible region is 'a line' only
Unbounded Solution — — — — — — — ^ — — — — — — — — — ^ — ^
Use graphical method to solve the following LPP
1. Maximise 2 6 x j + •4x2
Subject to -x 2 < 1
3xj - 2x 2 < 6;
xj,x2> 0
2. Maximise Z = 30bcx + 5<3x2

Subject to 2*! + x 2 > 7

X,[ + x9 > 6
2x! + 6x2 > 18
1 0, x 2 > 0
3. Maximise Z = 3x i + 2xc

Subject to ~2x ! + 3x 2 < 9

3x 2 - 2x 2 < 20
X! > 0; x 2 > 0
4. Minimise Z= ;3*, 1 — A X O
Subject to 2*! + x 2 < 1
X j + x2 > 5

* ! > 0,x2>0

5. Maximise Z-= 2
Subject to 4x, + 6x2 > 240
Xj +x 9 < 40
4xx + 3x 2 > 180
Xj, x 9 > 0
Remarks :

The solution the above problem (5) seems to be multiple optimal, but it is
unbounded.
Graphical Solution of LPP 79

Additional Problems
Solve graphically
1. Minimise Z = 30 xj +
Subject to Xj + x2 ^ 8
6xj + 4x2 > 12
5x} + 8x2 > 20

and JCJ, x2 > 0

2. Minimise Z= 2000x,+ 3000x 2
Subject to 6xj + 9x2 < 100
2x} + x2 < 20
x j, 0 [OU - MBA - Sep 98]

3. Maximise Z= 6IXJ + 7x2

Subject to 3xj + 9xj > 36

6xj + 2x 2 > 24

2xj + 2x 2 > 16
xlf 2 > 0 [0U - MBA - Nov. 94]
4. Maximise Z = x1 2 +X

Subject to 2xj +x2 = 5

3xj-x 2 = 6
x1? x2 unrestricted [0U - MBA Apr. 98, Sep. 98]
5. Provide graphical solution to
Minimise Z = 1000 x + 800 y
Subject to 6x+2y> 12
2x + 2y > 8
4x+ 12)i > 24
x,y> 0 [0U - MBA - Dec. 20000]
6. Maximise Z-x + y
Subject to 2x+3? = 5

x, y>0 [0U - MBA T Dec. 2000]

7. Maximise Z = 3xj + 5x2

Subject to + 4x2 < 9

>0 [0U - MBA - Sep. 98, Dec. 2000]

80 Operations Research : Theory and Practice

8. Minimise Z- 20xj + 14x2

Subject to x± + x2 < 18
5x1 + 3x 2 > 10
4xx + 6x2 = 20
xi > 0, x2 unrestricted - H8A - May 94]

9. Maximise Z = 2 0 x j + 40x 2
Subject to 36xj + 6x2 > 108
3xL + 12x 2 >36
20%! + 10%2^ 100
Xj, x 2 > 0 [0U - MBA Feb. 93]
10. Maximise Z = xx + 3x2
Subject to 3xj + 6x2 < 8
5*! + 2x 2 < 10
^ j , x2 > 0 [0U - MBA - Mar. 98]

11. Minimise Z=3x 1 +'5x 2

Subject to xj +x 2 = 200
x!<80; x 2 > 8 0
x1? x2 > 0 [0U - MBA - May 95]

12. Maximise Z = 2x1 + 5x2

Subject tc xx +4x2 < 24
3xi+x2<21
x t +x 2 < 9
xj , x2 > 0 [0U - MBA - May 95]

13. Minimise Z= 2500 *! + 3500 x2

Subject to SOxj + 60x2 > 2500
+ 60x2 > 3000
j 200 x2> 7000
Xj, x ? > 0 [0U - MBA -May 91, Dec. 94, Dec. 95]

14. Maximise Z = 3xj - x 2

Subject to 2xj + x 2 > 2

xx + 3x2 < 3

x2<4

Xj x2 > 0 [OU - MBA - May 92]

Graphical Solution of LPP 81

15. Maximise Z = 4xj + x2

Subject to 3*! + 2x2> 6
- 3^j - x 2 < - 4
4xj + 5x 2 = 10
Xi is unrestricted and x2 > 0 [0U - MBA - May 92]

16. A firm can produces two types of furniture viz. chairs and tables. The
contribution for chairs and tables is Rs. 200 and Rs. 300 respectively. Both are
processed on three machines Mh M2 andM 3 . The time required by each
product and total time available per week are given below.

Machine Chair Table Available Hours

M{ 3 3 36

M2 5 2 50
Ms 2 6 60

Formulate the LP and solve it, using graphical method determine optimal
production. [OU - MBA - Sep. 99]

17. What do you understand by graphic method? give its limitations use die
graphical method to find the amounts of factor^ and factor B to form a product
which must weight 50 kgs at a minimum total cost. At least 20 kgs of A and no
more than 40 kgs of B can be used. The cost of A is Rs. 100 per kg of B Rs. 25
per kg. [OU -MBA - March 99]

18. An animal feed company must produce 200 kgs of a mixture consisting of
ingredient Xj and X 2 . The ingredient Xj cost Rs. 8 per kg and X 2 Rs. 5 per kg.
Not more than 80 kg of Xj can be used and at lest 60 kg of X 2 must be used.
Find the minimum cost mixture. Use graphical method.
[OU - MBA - March 98]

19. Three magazines P.QjR are available for advertisement. Their exposure is
number of buyers multiplied by the subsequent number of advertisements.
Budget availability is Rs. One lakh manager feeds that P should have at least 2
adds each. Formulate LP model and solve it using graph.
[0U - MBA - July 2000]
82 Operations Research : Theory and Practice

20. Solve graphically

Maximise Z = 4x^ +5x^
Subject to xi + #2 - 1
- 2xi + X2 < 1
4*! + 2x 2 < 1
and x} x2>0 [0U - MBA - Sept. 98]

21. Use penalty method to minimize Z = 3X + 2.5 Y

Subject to 2X+4F>40
3X+2F>50
X, Y> 0 [JNTU (Mech./Prod.) Nov. 2002 (Set - 3 ),]

Review Questions
1. Discuss the merits and demerits of graphical solutions in operations research

2. What is meant by redundant constraint? Explain with examples.

3. Discuss different types of solutions with reference to graphical methods in linear

programming problems.

4. Discuss various types of optimal solutions to LPP through graphical methods.

How do you identify them

5. Discuss different types of feasible regions in graphical solutions to LPP

6. Discuss different types of constraints and explain how you represent them in
graphical solutions of LPP

7. How do you identify the following solution in graphical methods ?

(a) Multiple optimal solutions

(b) Infeasible solutions
(c) unbounded solution [0U - MBA - May 95, May 92]

8. What is an unbounded solution ? How is it different from infeasible solution

[0U - MBA - March, 98, March 99]

9. What is convex set? Explain with suitable examples. [0U - MBA - Nov. 94]

10. How are unrestricted variables treated in graphical solutions to LPP.

[0U- MBA -Dec. 95]
Graphical Solution of LPP 83

Objective Type Questions

1. The solution for an LPP with two exact constraints and no inequality constraints
will be
(a) infeasible (b) unbounded
(c) multiple optimal (d) unique point size solution

2. The feasible region will be type if one of the constraints is exact

type in a two variable LPP
(a) point (b) line
(c) area (d) volume

3. In graphical solutions, if x is unrestricted and y > 0, we get the solution in

(a) 1st quadrant (b) 1st & 2nd quadrant

(c) 1 st & 4th quadrant (d) any quadrant
4. In which of the following cases we do not maximise the objective function
(a) sales (b) profits
(c) contribution (d) costs

5. If every point of a line drawn with any two points in feasible region fall in the
feasible region the region is said to be
(a) convex (b) concave
(c) island (d) infeasible

6. In an LPP, if the values given to the variable satisfy conditions and constraints
but not the objective function, then the solution is
(a) feasible solution (b) optimal solution
(c) infeasible solution (d) we can not say

7. The feasible region in the form of a ring is

(a) convex (b) concave

(c) concavo-convex (d) convexo-concave

8. The number of basic feasible solutions in a feasible region will be

(a) finite (b) infinite

(c) zero (d) we can not say
84 Operations Research : Theory and Practice

9. T h e solution for max. Z = 1 0 x 1 x2 subject to 5xl + 3x2<30;

xY + 2x2 < 18; «!, x2 > 0 is

(a) unique (b) infeasible

(c) multiple optimal (d) unbounded

10. Which of the following constraints has no effect on solution space.

(a) exact constraint (b) availability constraint
(c) requirement constraint (d) redundant consu dint
11. There will be no solution in
(a) unbounded solutions (b) infeasible solutions
(c) multiple solutions (d) unique solutions
12. Which of the following solutions is independent of objective function
(a) unique (b) unbounded
(c) infeasible (d) multiple
Consider the graphical solution figure given below.

5001

5x,+3x 7 =1500

7x, + 9x? = 300

250
211.11A = 7x,+9x 7 =1900
200
•2xl +4x 2 = 1000

1/0(55.55)^300 500 ' '

42.83 D(271.43)

13. In the above graphical solution, which of the following is not redundant.
(a) lxx + 9x2 > 300 (b) 5*! + 3x2 < 1500
(c) 2x{ + 4x2 < 1000 (d) 9xx + 5x2 > 500

14. In the above figure, if objective function is to maximise 500 x± + 300 x2, then
we get
(a) infeasible solution (b) unbounded solution
(c) unique solution (d) multiple solution.
Graphical Solution of LPP 85

15. For the above figure which of the following objective function can give multiple
optimal solution
(a) 1170 xL + 1110 x2 (b) 1000 x{ + 600 x2
(c) 700 x L + 900 x2 (d) 1000 x! + 2000 x2

16. The get multiple optimal solutions in LPP which of the following is necessary
condition.
(a) objective function is perpendicular to one of the constraint
(b) any two constraints are parallel
(c) no two constraints are parallel
(d) objective function is parallel to one of the constraints.

17. The number of optimal solutions of LPP for a maximizing objective function
on unbounded feasible region is
(a) infeasible (b) unique
(c) multiple (d) infinite

18. The solution for the LPP : max. Z = 20 x± + 30 x2 subject to 5x^ + 6x2 < 15;
2*i + 3*2 - 200; x\> X2 - 0 m a y be solution
(a) multiple optimal (b) unbounded
(c) infeasible (d) unique

19. The solution for max. Z = Sx^ + 4x9 subject to x^ - x2 < 1, x± + x2 > 2 will be
(a) infeasible (b) unique
(c) multiple (d) unbounded

20. Which of the following is not a limitation in graphical solutions

(a) limited to two variables only
(b) accuracy is not possible
(c) infeasible solutions can not be predicted
(d) multiple solutions can not be predicted easily.

Fill in the Blanks

1. If a convex set of feasible solutions of an LPP is a bounded solution space then
its extreme points are said to be solution.
2. If a convex set of feasible solution of an LPP is a bounded feasible region then
it will have at least one solution.

3. solution is nothing to do with objective function of LPP.

86 Operations Research : Theory and Practice

4. constraint has no effect on feasible region

5. A solution is said to be if it satisfies the conditions,

constraints and objective function

6. Any set of values, which satisfy the conditions of LPP is a

7. Graphical solutions are limited to the problems with variables only

8. If multiple optimal solution exists to an LPP, at least one of the constraints will
have same slope as that of

9. If an LPP has infmite number of solutions, it is said to have solution

Answers
Objective Type Questions :
1. (d) 2.(b) Mb) 4. (d) 5. (a)
6. (a) 7. (b) 8. (a) 9. (c) 10. (d)
11. (b) 12. (c) 13.(d) 14. (c) 15 (c)
16. (d) 17. (d) 18. (c) 19. (d) 20 (c)
Fill in the Blanks :
1. basic feasible 2. optimal
3. infeasible 4. redundant
5. optimal 6. solution
7. two 8. objective function
9. unbounded
88 Operations Research : Theory and Practice

CHAPTER AT A GLANCE
Start

- Convert LP model into standard-form by

introducing slack/surplus/artificial variables.
- Decide the coefficient of these variables in.
objective function, (zero to slack/surplus and - M for
artifical variables)

No Convert min. form

^objective fiinctfoTi (•'•s to max. form by
\ v miximise s ^ multiplying with (-1)

YesJ
Rewrite objective function
using decision, slack/surplus,
artificial variables.

Obtain IBFS and identify basic and non-basic

variables, and solution values of basic variables.

Set-up initial simplex table

Compute Z . and Z - C . values to

all column's of simplex table.

Optimal solution is
obtained, write solution
values and find Z of A.

Select most negative value of

Z - C . and make it as key column

Find min. ratio by dividing

solution value by key column value

Are
the values o Solution is
min. ratio negative unbounded
or infinity

Select least value of min. ratio column and

make it as key row, and identify key element

Iterate to make key element as unity and other

elements of this column as zeros. Revise the
total table using appropriate rules of conversion.
Simplex Method 89

4,0 Introduction
Simplex algorithm (a step by step or iterative approach) was originally
proposed by G.B Dantzig in 1948.
It starts at a basic (the worst case) level of the problem.
At each step it projects the improvement in the objective function over its
previous step. Thus, the solution becomes optimum when no further improvement is
possible on the objective function.

4.1 Simplex Algorithm

The algorithm goes as follows :
Step 1: Formulation of LPP :
- Selection of decision variables
- Setting of objective function.
- Identification of constraint set
- Writing the conditions of variables.
Step 2 : Convert constraints into equality form.
- Add slack variable if constraint is < type.
- Subtract surplus and add an artificial variable if the constraint is > type.
- Add an artificial variable if constraint is exact (=) type.
Step 3 : Find, Initial Basic Feasible Solution (IBFS)
- If m non identical equations have n variables (m < n) including all decision,
slack / surplus and artificial variables, we get m number of variables basic
and (n - m) variables non basic (i.e, equated to zero).
- First make all decision variables (and surplus) as non basic i.e., equate to
zero to identity the IBFS.
- Find solution values for basic variables.

BV C
J
c
i Minimum Remarks
SV v
S& \ ratio
v
i
Solution values
Contribution of
basic variable

/ Most \
in objective

variables

minimum )
function

Basic

Key Element
ratio -*— Key row
(W 1 lofSV/keyi
\co. value/

Sum of products of CB and y{

zrCj Most negative value

Key column
90 Operations Research : Theory and Practice

Step 4 : Construct, Initial Simplex Tableau as given above with the following notations.
CB : Coefficient of basic variable in the objective functions
(or contribution of basic variable)
BV : Basic variable (from IBFS)
SV : Solution value (from IBFS)
Cj : Contribution of/'1 variable or coefficient of each variable (j ) in
objective function.
Z: - Cj : Net contribution.

Step 5 : Find 'out going' and 'incoming' variables.

- Find Z by summation of products of CB a n d ^ for each column.
- Computed 2? - C: value for each column
- To find key coloumn use most negative value of Z - C .
- Variable in key coloumn is 'in coming variable' or 'entering' variable.
- The variable of key row is 'out going' or 'exiting' variable.
- Find the minimum ratio of solution value to corresponding key coloumn
value to identify key row.
- The cross section of key column and key row is key element with which
the next iteration is carried out.

Step 6 : Re-write next tableau as per given set of rules.

- Replace the exiting variable from the basis with the entering variable along
with its coefficient (or contribution).
- You have to make key element as unity (i.e. 1) and other elements in the
key column as zeros.
- To make key element as unity, divide the whole key row by the key
element. This is supposed as the new TOAV in the place of key row in the
next iteration table.
- To find other rows of next iteration table, use this new row. By appropriate
adding or subtracting entire new row in the old rows, make other elements
of the key column as zeros.
- Accordingly the solution value gets improved.
Step 7 : Check whether all the values of Z: - Cj are positive. If all are positive, the optimal
solution is reached. Write the solution values and find Z opt .
(i.e., Z m a x or Z min as the case may be).
If Z - Cj values are still negative, again choose most negative among these and
go to step 5 and repeat the iteration till all the values of Z - C become positive.
Simplex Method 91

Start

Convert LP model into standard form by

introducing slack/surplus/artificial variables,
Decide the coefficient of these variables in
objective function, (zero to slack/surplus
and - M for artificial variables)

Convert minimum form

No to maximum form by
multiplying with (-1)
Yes
Rewrite objective function
using decision, slack/surplus,
artificial variables

Obtain 1BFS and identify basic and non-basic

variables, and solution values of basic variables

Set-up initial simplex table

Compute Zj and Zj - Cj values to

all columns of simplex table

Optimal solution is
obtained, write solution Stop
values and find Z of A
Yes
Select most negative value of
Zj - Cj and make it as key column

Find minimum ratio by dividing

solution value by key column value

Are
all the values of Solution is / _
minimum ratio negative unbounded I
or infinity
No
Select least value of minimum ratio column and
make it as key row, and identify key element

Literate to make key element as unity and other

elements of this column as zeros. Revise the
total table using appropriate rules of conversion.

FIGURE 4,1 : FLOW CHART OF SIMPLEX METHOD

92 Operations Research : Theory and Practice

The above algorithm is illustrated through the following numerical example.

ILLUSTRATION 1
Maximise Z = 6xt + 9x2
Subject to 2xj+2x2 < 24

6xj+2x2<60
and xpx2>0 by using simplex method. [JNTUMECH, ECE-1999/C]
Step 1: Foromulation :
The given problem is already formulated in canonical form.
Step 2 : Writing standard form by introducing slack variables;
(Given problem is in canonical form i.e., maximising objective function,
constraints < type and all non negative conditions)
Express given problem in the standard form i.e., objective function is to be
maximised (existing already), equality constraints (to be converted) and non-negative
conditions of variables (aready existing).
Standard form of the given problem is
Maximise Z= 6x^ + 9x% + Ojy + 0.52 + 0.53
2x^2*2 + 5^24 . . . . (1)
X1 "r OXo r 5o :— 1 1 .... (2)
6xj + 2^2 + 53 = 6 0 .... (3)
and xj >0, x2 > O.ij > 0,^2 > 0, ss > 0
Where 5^, 52 and 53 are slack variables added left side of equation (1), (2) and
(3) respectively so as to make the inequations (<) as equations. Also the condition of
these variables must be non-negative.
Step 3 : Now, we have three equations with five variables viz., x±, x%, 5j, 59 and 53 to get
the Initial Basic Feasible Solution (IBFS), we have to fix the values of any two
of these five variables. We can do this in bc ways, i.e., 10 types of solutions.
We choose x± and x2 at zero since the decision is to be made on them and we
start at most minimum possible value of x^ and x2 (x^ > 0, x^ > 0) and this is the most
worst case from where we iterate to improve the objective function to its maximum
value.
Thus IBFS is

[We now call * j , x2 as non basic variables]

=> *i = 24, 52 = 44, 53-60
[We now call sx, s2 and s$ as basic variables].

[Now, if the non-basic variables are brought into basis by the iteration, then the
optimal solution can be attained].
Simplex Method 93

Step 4 : Construct Initial Simplex Tableau - 1

INITIAL ITERATION TABLEAU
+
Basic
Max Z=

r rT t t Min Remarks
Variable Sol. Values Ratio

1 0 0

0 0

-60

/ - I -9 0 0 0

(OxO + O x l + OxO) (0x0 + 0 x 0 + 0 x 1 )

(0x2 + 0x5 + 0x2)

(0x1 + 0 x 0 + 0x0)
(0x2 + 0 x 1 + 0x6)
Step 5 : Identify key column (most negative Z - C.) and key row (min. ratio) and hence
the incoming and outgoing variable respectively.
-In coming Variable
ITERATION TABLEAU I

C
J 0 0 Min Remarks
BV Ratio
sv So

24 0 0 24/2

Key rowr
60 6 ; 2- ; -y o 0 1 60/2

0 0

Out going variable -6

Key/ coli
column
Step 6 : Re-write the next iteration tableau - II with the given set of rules :
[We can notice in iteration tableau that the values of body i.e.^, the coefficient matrix
for basic variables is in the form of a unit matrix. When s j , s2 and s 3 are Basic Variables, we
find unit matrix under their respective columns. When52 is replaced with x2 now, we should
make a unit matrix with sj, x2 and 53, the basic variables in the next iteration tableau. To
make this improvement, we follow the following rules.]
94 Operations Research : Theory and Practice

Rule 1 : Re-writting new row in the place of key row.

- The out going variable (s2) is replaced by incoming variable (x2) in the
basis. Obviously the corresponding contribution i.e., CB ( '0' in old row)
changes (as 9 in new row).
- Now, key element ( 5 in old row) is to transformed as unity. Therefore the
total row is divided by 5.
N 1
Thus /?2 ~* "7^2 [R ea d as New /? 2 is changed as one fifth of old i?2]

Rule 2 : Re-writing other new rows.

- Other basic variables and their contribution (C^) remain same i.e., figures
in first two columns are unaltered except for the replacement as given in
rule 1 discussed above..
- The other values in the key column must possess zeros therefore, subtract
or add R^ appropriate number of times so as to get zeros.
- Thus entire row is re-written by adding or subtracting the new row found
from rule 1, W number of times if m is the vaue in old row.
- Thus here Rx is transformed as R^ -> R^- 2R^ T~
(Read as new R j is changed as old R j minus two times new i?2)
we subtract tms \ ^.—— —-—^^^
'V two times to ) /we have to make this zeroN rule
It is illustrated here below : V^makezero </ y$o mbtract key row^timgs/
Old Rx is 0 S{

New
N ewiR
?9 is
2's 9 *2 —- - - 1- (T
(obtained from rule 1)
44 1 1
New/?! is 0 S l 24 - 2 x — 2- 2x- 2 - 2 x 1 1 - 2 x 0 0 - 2 x - 0 - 2 x 0
5 5 o
32 8 2
!
T 5 ° -5 °
Similarly, R^ -> R$ - 2 R%
9 v 44 2x1 1
N e w / ? , is 0 S$ 6 0 - — - — 6 - ^ - — 2 - 2 x 1 0 - 2 x 0 0 - 2 x - 1 - 2 x 0
5 5 5
Simplex Method 95

These are written in tableau - II given below and the above steps are repeated.
i coming Variable
ITERATION TABLEAU U

0 Min
B.V Remarks
Ratio
sv
'W^:m:i-:
Key element
44 .••.-.•I1-:- , 1 44/5
0
5 "5 1/5
212 2 212/5
0 0
5 ~ 5 28/5

Out going variable

^A
zrc.
\
Key column
w 0 x 0 + 9 x 1 + 0
8 n 1 28 0x0+9x0+0x1
0 x —+ 9 x —+ Ox —
3 5 5 0x1+9x0+0x0
ITERATION TABLEAU
C 6 9 0 0 0 Min
B.V J Remarks
x2 s2 Ratio
SV *i *3

6 x{ 32 5 8 5 2 5
0x5/8
TX8 5X8 ~5X8
=0
= 4 =1 =0 = 5/8 = -1/4

9 x2 44 1 . 1 1
1-^x0 5 8 5 5l 4j 0-^x0
T"5x4 5 5
1 - - 5 l

=8 x 1=0 =1 = =0
~~8 4
212 28 28 28 2 28/"-H
0
0-fx0 0 ^ 1-fxO
— - T x 4 T"T x l =0
° ~ 5X8
= -7/2
5 5 4
= 1V ; =1
= 20 =0
6 x 1 6 x 0 6x5/8 6 ( - lA) 6 x 0
h + 9 x 0
+ 0 x 0
+ 9 x 0
+ 0 x 0
+ 9 ( - 1/8)
+ 0(-7/2)
+ 9 (V-i)
+ 0 x 1
+ 9 x 0
+ 0 x 1

ZJ-CJ r6 - fi
6=n0 9-9=0 f-° 21
I"0 3
°-°
8 4

Iteration tableau - III is re-written with simplified values.

96 Operations Research : Theory and Practice

C
i 6 9 0 0 0 Min
B.V Remarks
S2 53
Ratio
SV x} xt S{

6 x
i 4 . o f -i o
9 x2 8 0 , -i I 0

0 20 0 0 - | 1 1

6 9
f ! °
21 3
h
91 ^
0 All positive, so Z max = 96
° T ! °
z c
r j

As the values of Z - C are positive for all the variables, the optimisation is
reached.

Therefore, the solution is

= 8 I Basic

Non-basic

=6x4+9x8+0x0+0x0+0x20

= 24 + 72 = 96

Remarks :

(i) From the above solution values, we can understand that 20 units of third resource are
left unutilised while first and second resources are fully utilized.

(ii) Students are advised not to express the fractions in the decimal form, since these will be
normally simpler calculations. If they are expressed in decimal form, it viay affect the
accuracy.
Note : (i) While calculating minimum ratio, ifyou get negative value or infinity ignore them.
(ii) A minimisation objective function can be changed to maximisation by multiplying with (- 1). You
can observe these in the illustrative example to follow.
Simplex Method 97

Let us consider an example with minimisation objective function.

ILLUSTRATION 2

Minimise Z=2x2 6x2 + 4x3

Subject to 3xjt - x2+3xs<7
-2xj+4x2<12
- 4x2+3x2+5x3<10
xp xPx3>0
Solution :
Step 1 : Formulation : It is already formulated but it is not in the canonical form with
reference to its objective function.
Therefore we convert into canonical form.
Since, Minimising Z = (-1) Maximising Z
(Mini - Max theorem),
we rewrite the problem as,
Maximise Z = - 2 Xj + 6 x2 - 4 x3
Subject to 3x1 - x2 + 3x3 < 7
- 2 x 1 + 4 x 2 < 12
- 4xj + 3x 2 + 5x 3 < 10

Step 2 : To write in standard form by introducing slack variables,

3 x± - x 2 + 3 x3 + Sy = 7
-2xl+4x2 + S2 = l2
- 4 x{ + 3 x2 + 5 ;c3 + S3 = 10
and maximise Z = - 2 #j + 6 x 9 - 4 x 3 + O.Sj + 0.59 + 0.5 3
x±, x2, ^3, Sj, 5 9 , 5g > 0
Step 3 : In the above problem, we have three (3) equations and six (6) variables.
Therefore we fix three of them in basis by assigning zero value to the other
three (non- basic) variables.
(6 5 4 ^
This can be done in 6C or ' ' = 20 ways, leading to 20 different types of
solutions. Of them, we choose the initial basic feasible solution (IBFS) as the decision
variables at non - basis i.e., assuming zero values. [This indicates the unutilized
resources when the production is nil or stand still].

Thus when x{ = 0, x2 = 0 8c x3 = 0 the solution values (SV) for Si = 7, S2 - 12

and S3 = 10
98 Operations Research : Theory and Practice

Step 4 & 5 : Setting initial simplex tableau and identifying key column, key row and hence
key element. Also locate in coming and out going variables.
ITERATION TABLEAU -1
i coming Variable

C'B B.V
Cj -2
V A
— Tt
0 0 0 Min
Ratio
Remarks
sv *i
7
0 «i 1 3 -1 3 1 0 0 —-~-V€
*
- } (Ignore)
12 0 0 ".;.:. Keyyow v,;
Key element
: , ' - • •

0 10 -4 3 5 0 0 1 10
T>3
h 0 o 0 0 0 0
Out going variable
zr C
J 2 4 0 0 0

Key column
* Ignore if minimum ratio is negative or infinity.
Step 6: The tableau - I is iterated to give tableau - II, according to given remarks in
tableau - II
(Find R^ by ~ R% and then R^ and R$ using R$).

ITERATSOW - TABLEAU - H
i coming Variable

C
J -2 -4 0 0 0 Min
B.V Remarks
Ratio
sv
10 i 3^S
-Key element

i ° (ignore)

-I o . -! • (ignore)
Out going variable Z- - 3 0 0 3/2 0

-I-"" 0 0 3/2 0

Key column
Simplex Method 99

ITERATION TABLEAU -11! (first find itf hy~ ft? m\d then R$ and Rf)

- 9 6 -4 0 0 0
c
C
B.V J Min
Remarks
sv x2 s Ratio
Xj
\

-2 x
\ 4 1 0 6/5 2/5 1/10 0

6 *2 5 0 1 3/5 1/5 3/10 0 * ^+l <

1
0 11 0 0 8 1 1
~2
h -2 6 6/5 2/5 8/5 0
z
r C
j
0 0 26/5 2/5 8/5 0 Z min = - 2 2
All the values of Z - C- are positive, therefore optimality is reached.
Solultion is
Xj = 4, #2 = 5, X3 = 0
(Also 5j = 0 , 5 2 = 0, 5 3 = 11 i.e., unutilized resource)
z
max = - 2 x 4 + 6 x 5 - 4 x 0 + 0 x 0 + 0 x 0 + 0 x 1 1
= - 8 + 30 - 22
7
(at xj = 4 , x 2 = 5, X9, = 0)
Now, let us consider a case, with minimisation objective and mixed type of constraints
(one of the cosntraints is > type)
ILLUSTRATION 3

Min Z= Xj-x2-2x3
S.t 3xj+ x2+3x3<7
-Xj+2x2>-6
4xj+3x2+5x3<10
xp xPx3>0
Solution:
Step 1: Formulation : Already formulated but not in canonical form.
Converting into canonical form.
Max Z =-Xj + #2 + 2 x3
S.t. 3x1 + x2 + 3 x 3 < 7
xx - 2x2 < 6
4xj + 3x2 + 5x 3 < 10
Xj, x 2 X3 > 0
(Here, obj. function and second constraint are multiplied by - 1)
100 Operations Research : Theory and Practice

Step 2 : Writing in standard form by introducing slack variables.

3*! + x2 + 3x$ + Sx = 7

Xj - 2^2 + ^2 = 6

Ax^ + 3^2 + 5x3 + S 3 =10

And obj. fun : is max Z = - x^ + x2 + 2 x3 + 05j + O.S2 + O.S3

xj, x2, x3 > 0, Sj, S2, ^3 - ^

5^3; Finding IBFS

Tliere are three equations and six variables

.*. Three of them are kept in basis and three in non basis ( we get 6C types of
solutions)

IBFS : non basic x± = 0, x2 = 0, x3 = 0

basic 5] =7, S2 = 6, 5 3 = 10

(fs5 5; Write initial tableau, identify key column, key row & key element, and in coming
& out going variables
ITERATION TABLEAU • i :
llll

-1 +1 0 0 0 Min
C B.V
Ratio
Remarks
s.v.
7
0 Si 1 3 + 1 1 0 0
3 If Min ratio
is -ve or «>, it
is ignored
0 s2 6 -1 -2 0 I 0 0
— = oo

(ignore)
•Key element
' ;•'•./ . • ":"
:
•'•:•" • ^ • ' v " '
k

k-mi • > , •
.:•?. . - " • 5
• - "
o; !
kl-:---5 - : k :V-.'••:
Key row

h 0 0 • - . • • • " " •
0 0 0

+1 - 1 0 0 0

Key column
Simplex Method 101
Step 6 : Iterate Tableau - 1 , according to given remarks in tableau - II
^ITERATION -TABLEAU -II:

C
J - 1 1 2 0 0 0 Min
C B.V Ratio
Remarks
s.v.
0 1
i -f » • • - i
0 s2 6 1 - 2 0 0 1 0 /?2^ —> i?2 ( n o change)

2 2
I ! > » » i
Z
j ! ! » • • ! Since Z - C•> 0 V x- and 5y
we have Z m a x = 4
Z
i~CJ
f 1 • » • I
As Zj - C: values are positive for all the variables, optimality is reached. Optimal
solution is
xy = 0, x 2 = 0, x% = 2

and Zmax =-0 2x2 + 0x1+0x6 + 0x0

=4
Zmin = - 4 at #! = 0, x2 = 0 and x 3 = 2
Remarks :

Converting a 'greater than or equal to' (>) type constraint into 'less than or
equal to* (<) type constraint is not applicable to all the cases. Particularly,
(i) When all the constraints are 'greater than or equal to' type, this method is not
suitable, and

(ii) This is applicable when right hand side of inequation is negative.

In all such cases we may have to choose either BIG M or 2 phase method to be
discussed in the sections to follow.

4.2 Some Special Cases

We have learnt how we get multiple optimal solutions, infeasible solutions and
unbounded solutions by graphical solution method. Let us now observe how we
identify such solutions in simplex method.
102 Operations Research : Theory and Practice

4.2.1 Unbounded Solution

In some cases if the value of a variable is increased indefinitely, the constraints
are not violated. This indicates that the feasible region is unbounded at least in one
direction. Therefore, the objective function value can be increased indefinitely. This
means that problem has been poorlv formulated or concieved.
In simplex method, this can be noticed ifZj- Cj value is negative to a variable (entering)
which is notified as key column and the ratio of solution value to key column value is
either negative or infinity (both are to be ignored) to all the variables. This indicates that
no variable is ready to leave the basis, though a variable is ready to enter. We cannot
proceed further and the solution is unbounded or not finite.
This is illustrated through the following example.
ILLUSTRATION 4

Maximise Z=4x2+ 3x2

Subject to x2<5

xpx2>0 [IGNOU-MBA-1997
BRAOU-MBA-1998]

Let us introduce slack variables to express inequalities;

Maximise Z = 4xj + 3x2 + 0.5] + 0-s2
Subject to x± +.Si = 5

i 2 0 , Si,$2 ^ 0

IBFS 2^ = 0, x2 = 0 (Non basic)

s± = 5, -s2 = 8 (basic)
With usual steps it is solved through the following iterations.
ITERATIONS TABLEAU E
-Entering Variable

Key row
Simplex Method 103
ITERATION TABLEAU II
R1? -^ R i ; -Entering Variable

0
BV Min ratio Remarks
sv s2
to be ignored

to be ignored

No variable is ready to leave the basis

z c Solution is UNBOUNDED
r i

Key column

From the above tableau - II, it is clear that x2 is entering variable into the basis
(key column variable whose Z: - C: is negative) but no variable is ready to leave since
the ratio of solution vaue to key column value is infinity for R± and negative for /? 2
both of which are to be ignored. Thus we cannot proceed further because key row
(leaving variable) cannot be found. Thus, this problem yields no finite solution or in
other words, an unbounded solution.

42,2 Multiple Optimal Solutions

When the objective function is parallel to one of the constraints, the multiple
optimal solutions may exist. As we have seen from graphical solutions, that the optimal
solution exists at the extreme point on the feasible region, the multiple optimal
solutions will be noticed on at least two points of the binding constraint parallel to
that of objective function. Thus in simplex method also at least two solution can be
found.
This alternate optima is identified in simplex method by using the following
principle.

After reaching the optimality, if at least one of the non-basic (decision) variables
possess a zero value in Z: - Cj, the multiple optimal solutions exist.

This is illustrated through the following numerical example.

ILLUSTRATiOiy 5 — _ —
Maximise Z=3x2+ 6x2
Subject to x2+ x2<5
Xj+2x2<6
xpx2>0
104 Operations Research : Theory and Practice

solution :
Converting inequalities into equations, we get
Max Z = 3*2 + 6x2 + O.S{ + O.S2
Subject to XJ+X9 + 5J = 5
V -L 9 -V - U ^ — ft

Aj T A X9 T~ J O — U

xx 5 S >0
IBFS : Sj = 5, S2 = 6 (Basic)
#! = (), x2 = 0 (Non basic)
ITERATION TABLEAU - 1 :
- Entering Variable

Leaving variable
Key column
ITERATION TABLEAU - 1 ! : Entering Variable

Since Z - C • > 0 V all variables,

Optimal solution is 2 m a x = 1 8

Key Column

From the above Iteration Tableau - II the optimality is already reached since,
- Cj> 0 i.e., positive for all the variables. Therefore the solution is

and Z m a x = 3 x 0 + 6 x 3 =18

However, the value of Z- - C.- = 0 for the non basic decision variable xx in tableau
- II. This indicates alternative optimal solution. Therefore choosing first column as
key column, if we further iterate, we get the following tableau - III
Simplex Method 105
STERATIO&J TABLEAU - H i :

C
J 3 6 0 0
BV Min ratio Remarks
SV #j ^2 Oj 09

3 x
\ 4 1 0 2 - 1 5

6 x2 1 0 1 - 1 1 3 piV __. p 0 ^ TDN

Z
J 3 6 0 3 Z • - C • > 0 for all variable and hence

z c 0 0 0 3 Optimal solution is Z m a x = 1 8
r j

This tableau - III yields another solution as,

#1=4, #2=1
Z max = 3 x 4 + 6 x 1 = 18

We can observe that Z - C value will be always zero for basic variables in any
simplex tableau [observe for S} 8c S9 in Tableau - I , for Si 8c x9 in Tableau - II and
for %i 8c x9 in tableau - III they have Z - C values zero .] However, in Tableau - II?
apart from basic variables Si and x2 , the other variable xj is also showing Z: - Cj~ 0.
This indicates that xj is also worth coming into basis, as it is behaving similar to any
basic variables. Thus it provides an alternate optimal solution.

However, if all the basic variables are not replaced by decision variables it does
not mean any multiple solution unless this condition (i.e., Z - & = 0) is obeyed by any
non - basic decision variable.

4*2.3 Unique Solution

The problem solved in illustration 1 to 3 are the examples of this type of
solution. Further, one more example is shown below.
ILLUSTRATION 6
Mayukha & company produces two types of toys. Labour time required for first
type is twice that of the seconds. If the first type alone is produced, the capacity is
sufficient to produce 1000 numbers per day. The demand potential is 300 and 500
toys. The profits per toy are Rs. 10 and Rs.8 respectively. Formulate the L.P
problem and solve it. [J N T U (Mech/Prod.) 2001,2003]

Solution:
Step 1 : Formulation :
Variables: Let the no. of toys produced in first type = Xj
and the no. of toys produced in 2nd type = x2
106 Operations Research : Theory and Practice

Objective function: Profit on each first type toy = Rs.10

Profit on xj no. of 1st type toy = 10 xx
Profit on each second type toy = Rs.8
Profit on x2 no. of 2nd type toy = 8 x2
Total profit = 10 xx + 8 x2
.: Objective function: maximize Z = 10xj + 8x 2
Constraints:

(i) Labour time constraint: As labour time of first toy is twice that of second, if xj of
first type is produced, 2 x2 of second can be produced.

Total xj + 2x2 < 1000 (constraints on labour time)

(ii) Demand potential constraints:

x2 < 300, x2 < 500 (constraint on demand)

Since negative production is not possible

x^ , x 2 > 0
Summary Max Z = 10 x i + 8 x2
subject to xi + 2 x 2 < 1000
xx <300
x2 <500
Xj , x 9 > 0
5fe/> 2 : Conversion into equations by introducing slack variables (standard form)
Max Z = 10 xL + 8 x2 + 0-Sl + 0-S2 + 0-S3
x 1 + 2x 2 + S 1 = 1000,
xl + S2 = 300,
x 2 + S3 = 500
Xj , X2 , Sy , S 2 , ^3> 0

Step 3 : IBFS : Total variables = 5

Total equations = 3
Basic variables = 3
i.e., Sx = 1000, S2 = 300 and 5 3 = 500
Non basic variables = 5 - 3 = 2
i.e., X! = 0 a n d x 2 = 0
No. of basic solutions = 5C2 or 5C3 i.e. 10
Simplex Method 107
ITERATIOiy TABLE I :

C
J
10 8 0 0 0
o. B V
SV 52
Min Ratio Remarks

0 1000 2 1 0 0 1000

•
Si

0 300 0 0 1 0 300 «—
0 500 0 1 0 0 1 oo (Ignore)

Z
i
0 0 0 0 0

ZfCj -10 -8 0 0 0

ITERATION TABLE - M :
C 10 8 0 0 0
J
ca BV
SV Xj x2 oj ^2 *^3
Min Ratio Remarks

0 700 o [Tl i -l o 350 <- iTi

l
—mi—Ao
l
N
10 xv 300 1 0 0 1 0 oo (Ignore)

0 500 0 1 0 0 1 500
Zj 10 0 0 10 0
z c
ri 0 -8 0 10 0

ITERATION TABLE - i l l :

C 10 8 0 0 0
J
C B V Min Ratio Remarks
SV x
\ Si s2 s.
8 H 350 0 1 1/2 -1/2 0

10 X, 300 1 0 0 1 0 R»->Rl
0 s, 150 0 0 -1/2 + 1/2 1 Rt^Rl-RNi
Z
J
10 8 4 6 0 xx = 300
2 ma x-5800
0 0 4 6 0 x2 = 350

Since Z - C- > 0 V variable, optimal solution is arrived.

xx = 300,
108 Operations Research : Theory and Practice

x2 = 350
z
max= 10x300 + 8x350
= 3000 + 2800
= Rs. 5800/-
(unutilized resources S3 = 150 units).

4,2.4 Infeasible Solution

There may not exist any solution to certain LPP. This in LPP jargon is said to
be infeasible solution. In this type of solution, there exists no feasible region (refer
graphical solutions). We do not get any infeasible solution with all constraints as * less
than or equal to type'.
This will be explained later. However, to get an awareness, consider the
following example.
Maximise Z = 2 xx + 3 x2
Subject to xx < 5
xx - x 2 > 10
xx>0, x2 > 0
Solution ; Converting the problem into canonical form,
Max, Z =2xx + 3x 2

Subject to xi < 5

- x j + x2 < - 10

xx > 0 , x9> 0
Now introduce slack variables, to get equations
Xj + Sx - 5

^ 01
IBFS : ^} Non basic
x2 = 0 j
Si = 5 1
o in ^ Basic
^2 = - 10

From the condition 5 2 > 0, S2 can not take value - 10, therefore no IBFS can
exist with such less than or equal to type constraints.

So also, this method of converting' > ' type consraint to' <' type with a negative
sign can not yield any result and not suitable in such cases, as said earlier.
Simplex Method 109

Practice Problems
1. Max. Z=2x 1 + 5x 2
O.I. A Xi ~r Xo

2x2< 460
#j, x2 > 0 [JNTU (CSE) 97/sj
;4n$ujer; Unique solution : x± = 100, #9 = 230 and Max Z = 1350
2. A firm manufacturers two products in three departments. Product/1 contributes
Rs. 5/- per unit and requires 5 hrs. in dept. M, 5 hrs. in dept. N and one hour
in dept. P. Product B contributes Rs. 10/- per unit and requires 8 hrs. in dept.
M, 3 hrs in dept N and 8 hrs in dept P. Capacities for departments M,N,P are
48 hours per week. Find out optimal product mix using Simplex model.
[JNTU (CSE) 98]

Answer : Unique solution : x^ = 0 x% = 6, Z m a x = 60

(Refer illustration 3 of chapter 2 for formulation)
3. Formulate the following problem as an LP Problem:
A firm engaged in producing 2 models X± , X2 performs 3 operations -
Painting, Assembly and Testing. The relevant data are as follows :
Hours Required for each unit
Unit Sale Price
Model Assembly Painting Testing
Model Xx Rs. 50 L0 0.2 0.0
Model X2 Rs. 80 1.5 0.2 0.1
Total number of hours available are :
For Assembly 600
For Painting 100
For Testing 30 [JNTU (Mech.) 98/S]
Determine weekly production schedule to maximise revenue.
Answer : Unique solution : x±= 150, x% ~ 300
Max Z = 31,500
(Refer illustration 2 of chapter 2 for formulation)
4. Use simplex method to solve the following LP problem and verify by graph.
Maximize Z = 3 xx + 4 x2
Subject to the constraints :
2 xL + x2 < 40 ; 2 xY + 5x2 < 180 ; xY , x2 > 0 [JNTU (Mech) 98, 0U - MBA Jully 2000]

Answer : Unique solution Xj = 2.5 x 2 = 35

Z m a x = 147.5
(Refer illustration 4 of chapter 3 for graphical solution)
110 Operations Research : Theory and Practice

5. Find the minimum value

Subject to the constraints
^ 10
+ X2 -

%i , x2 > 0
Solve graphically.
Use simplex method to verify its graphical solution. [JNTU (ECE) 97]
Answer; Unique solution x± = 2, x2 = 0 ; Min - - 2
(Refer illustration 4 of chapter 3 for graphical solution)
6. Using simplex method,
Maximize Z = 5xi + 4x2
Subject to constraints 4 xx + 5 x2 < 10

3 x 2 < 12

* 1 >X2 - ° [JNTU (ECE) 97]

15 8 653
; Unqiue solution x± -~r , ^ = ~z\ MaxZ = - ~ -
7. Put the following problem into a standard form of L.P.P :
Maximise f(x) = 2xj + 6x2
subject to - x j + ^ 2 - 1
2x1+*2<2
x t > 0 , x2 > 0
Hence solve the problem by graphical method or otherwise. [JNTU (EEE) 97/S]

Answer; Unique solution x^ - — ^ 2 - V ! Zmax = —

3 O O

8. A company wants to purchase at most 180 units of a product. There are two
types of the product, Mx a n d M 2 available. Mx occupies 2/? 3 , cost $12 and the
company makes a profit of $3. Af3 occupies 3/2 3 , cost $15 and the company
makes a profit of $4. If the budget is $15,000 and the warehouse has 3000 fiz
for product.
(i) Set up the problem as a linear programming problem.
(ii) Solve the problem using simplex method. [JNTU (EEE) 98]

Answer : (ii) Unique solution xx = 0, x2 = 100 ; Z m a x = $ 4000

(for (i), Refer illustration 7 of chapter 2)
Simplex Method 111

9. Max. Z = 45x 1 + 70x 2

S.t. 2 * 1 + 3 * 2 < 70
3x } +5x2 < 110
xx+4x2 < 100
xl > 0, x 2 > 0

Answer: Unique solution x± = 20, x2=10 Z m a x = 1600

10. Max Z = 8tfI + 6 * 2

S.t. 2 x 1 + x 2 ^ 72
^! + 2 x 2 < 48
xx > 0, x2 > 0 jJNTU. Mech. 93]

^4n5a;er : Unique solution Xy = 32, ^9 = 8 Z m a x = 304

(Refer illustration 1 of chapter 3 for graphical solution)

11. Max Z = 5 ^ + 6x2 + 8x 3

S.t. 2 * 1 t * 2 + 2x3 < 50

xj + 6 x2 + 2 x% < 50

^j + 2x 2 + %3 < 26

Xy, x2, x^ > 0

Answer: Unique Solution xj = 0, x2 = 0, x 3 = 25; Max Z = 200

12. Max Z = 2 * ! + 4 x 2 + 2*3

S.t. 2 JCJ + x 9 - x 3 < 2

- 2 xj +x 2 - 5 x3 > - 6

4 x j + A:2 + ,\'3 < 6

x±, x2, x 3 > 0

Answer: Unique solution, ^ ^ 0 , x2 = 4, x 3 = 2 ; Max Z = 20

13. Maximise Z : 3 ^ + 2*2 ;

S.t constraints 2xx + x2 < 2

3xj + 4x2 > 12 and

x
\> x2>0 [JNTU . Mech/Prod/Chem. 2001/S]
112 Operations Research : Theory and Practice

14. A small paint factory produces three types of paints as follows :

Paint Production in kg/day Profit, units /kg
1 X 10
2 Y 4
3 Z 1

Due to limitations on the processing machinary, the following constraints are

to be met with :
X+Y<5
2X+F+Z<20
Using the simplex tableau, find out the product values of all the three paints,
which will maximise the piofit; and the value of the profit. [JNTU - CSE/E Comp. E 2001]
Unbounded Solutions — ^ — — — — — ^ — ^ ^ — i i ^ — — — —
1. Solve by simplex method
Max Z=3x!+4x 2
S.t. xj - x 2 < 1
- X j +X 2 < 2
xlfx2 > 0 [JNTU (CSE) 2001]

Answer: Unbounded solution

2. Max Z = 6 x 1 + 4 x 2
S.t. Xi - Xo < 1

x} > 0, x2 > 0
Answer : Unbounded or no finite solution

S.t. 2xj ~ 3x 2 > - 9

3x1-2x2<20
x i , X9 > 0
Answer ; Unbounded or no finite solution

4. Maximise Z = - 2 x± + 3 x2
x, < 5

3 2 "
Xj , x 2 S 0
Simplex Method 113

3
Hint; In final table at xj = 5, x2 = 9 and Max Z = 15 but Z - C• = - —.
All the elements in key column (second column) are negative.
Hence solution is unbounded
5. (a) Min Z = - 6 x^ + 2x2

S.t. xi - — x2 < 1; xj < 4

and JCJ , x2 > 0
(b) What is your comment if obj. function is Min Z = 6 x^ - 2 x9
4nsjwer ; (a) Unbounded or no finite solution.
(b) At Xj = 4, x2 = 6, Max Z = 12, But all the elements in second column
are negative. The feasible region is unbounded but the optimal solution
is bounded.
6. Show that the LPP
Max Z = 4 xy + x2 + 3 x 3 + 5 x4
Subject to 4 xj - 6 x2 - 5 x3 - 4 x4 > - 20,
x 4 < 10 ,

x1? x2, x3, x 4 > 0

Has an unbounded solution
Multiple Optimal Solutions — ^ ^ ^ ^ — ^ ^ ^

S.t. <48
2x! <9
x2 <6
xj >0, x2 > 0
Multiple optimal solution 4.5, x2 = 2, Z m a x = 24
1.5, x 2 - 6 , Z l l l a x =24
2. Max Z = 0.4 *! +0.2*2
S.t 2x L + x2 <72
xj + 2 %2 < 48
Xj , x 2 > 0

Answer : Multiple optimal solution 32, x2 = 8 or

36, x2 = 0 for Max Z= 14.4
114 Operations Research : Theory and Practice

3. Max Z=
S.t. x+y <40
x <20
15x+ 12)><300
x > 0, y > 0
Answer : Multiple optimal solution x = 20, y = 0 or
* = 0 , ^ = 20, Z m a x =100

4. Max Z = 6*! + 22 3

S.t. 2x{ + 3x2 + x3 <28

3 xj + x2 + 2 x3 < 24
< 35

; Multiple optimal solution x± = 8, x2 = 0, #3 = 0; Max Z = 48

5. Max Z = 6x! + 3x 2
S.t 2x 1 + 3 x 2 < 8
xj +x 2 < 6

Xj, X2 > 0
Answer : Multiple optimal solution xj = 4, x2 = 0
5
or xj = •—, x2 = 3 and Max Z = 24
6. Show that the LPP Max Z = 0Axx+x2
S.t 2xl+x2<5O;2xl + 5x2

Has multiple optimal solutions.

Additional Problems
1. Max Z = xj — x 2 + 3 X3
S.t. Xj + x 2 + X3 < 10
2xl-x^t <2
2 X 2 + 3 X 3 <0
xx,x2,xs >0
Simplex Method 115

2. Use simplex to maximise Z- 5 x^ + 2 x2 + 3 x3

S.t. xx + 2x : 2 + 2 x 3 + x 4 = 8

3 xj + 4 x 9 + x 3 + X5 = 7
Xj > 0 where 7 = 1, 2, 3 , 4 , 5

3. Max Z = 4 xj + 3 x 2 + 4x3 + 6x4

S.t acj + 2 x 2 + 2 x 3 + 4 x 4 < 80

;j + 2 x 3 + x 4 < 60

3x 1 + 3 x 2 + x 3 + x^ 80

Xj<0 where 7 = 1 to 4
4. Max Z = 4 Xi + 5x 2 + 9 x 3 + 11A

S.t. x1 + < 15

7x1 + 5x 2 + 3 x 3 + 2x 4 < 120

+ 5x 2 + 10 x 3 + 15 x 4 < 100
x:>0,j= 1 to 4
5. M a x Z ^ - 3 x 2 + 2* 3
S.t. 3 * : , - * 2 + 2x 3 < 7
^! - 2 x2 > - 6
- 4 x j + 3x 2 + 8x 3 < 10

X1 , X o , XQ >0
6. Max Z = 5 Xj + 4 x2

S.t. 4x! + 5x2 < 10

3 Xj + 2 x9 < 9
8 xj + 3 x2 < 12
*l > 0
x2 0

7. M a x Z ^ + 2X9 + 3 x 3
S.t. XtHh 2 x2 + 3 x3 10

<5

xlt: >0
116 Operations Research : Theory and Practice

Max Z = 2 %i + 4 x2 + x§ + x4
S.t xx + 3x 2 + x4 < 4
2JC 1 +A- 2 <3

x9 + 4 x 3 + x4 < 3

x j , X9, X3, X4 > 0

9. Max 15 xj + 6 x2 + 9 x 3 + 2 x4
S.t. 2 x j + x 2 + 5 x 3 + 0.6x 4 < 10
3XJ+X2 + 3X3 + 0.25X4 < 12
7 x j + x 4 <35
x7->0 where j= 1, 2, 3,4
10. MaxZ = 2x 1 + 5x 2 + 6x 3
S.t. 5x1 + 6 x 2 - x 3 < 3
- 2 %\ + x9 + 4 x3 < 4
Xy — 5 x2 + 3 X3 < 1
- 3 «!-3^2+7^3 <6
and xj , x9 , x3 > 0
11. Max Z = 3 x 1 + 4x9 + x3 + 7x 4
S.t. 8x{ + 3x 2 + 4 x3 + x4 <7
2xj + 6x 9 + x3+5x 4 < 3
x\ + 4x 2 + 5 x3 + 2x 4 < 8
and x lt x2, x3, x4 > 0
12. Max Z- 3 Xi + 4x 9 + x3
S.t. x1 + 2x 2 + 3x 3 <90
2XJ+X2 + X3 < 60
3 x ^ x 2 + 2x3 <80
and xj, x2, x3 >0
13. Max Z =.2 xj - 4x 2 + 4 X3
S.t. x1 + 5 x 9 - 5 x 3 <8
- 6 xi + 7x 2 - 7 X3 < 3
and *i>0, x3>0
Simplex Method 117

14. M a x Z = 5 x 1
S.t. <2
* 2 < 10
x2 < 12
and x2, - 0
15. Min Z = #! - 3 x2
S.t. - x 2 + 2x 3 < 7
2x 1 + 4x 2 < 12
3X2 + 8X3 < 10
and [IAS-Main 93]

16. Max Z =
S t 3x — ^x + 2x < 15
<3
x
l ~ X2 + ;3C
3 <4
xlfx2,xs >0 [0U - MBA Apr. 99]

17. MaxZ=23x! + 32x2

S.t. IOXJ + 6x2 <2500
5xj + 10x2 <200
<500
xlf*g >0 [OU - MBA - Apr 94, Sep. 99]

18. Solve : Max Z = 400*! + 30(k 2 + 2OO*3

S.t. 40*! + 20*2 + 3Cbc3 <600
<400
30*, + 2(k 2 + 10x 3 <80
Xi y X o , XQ >0 [OU - MBA 99/S)
19. Use simplex and verify by graph
Max Z = x1 + 3x2
S.t. Sxy + 6x 2 < 8
5xj + 2x 2 < 10
>0 [OU - MBA - Apr 98]
118 Operations Research : Theory and Practice

43 Simplex Method: Using Surplus & Artificial Variables

So far we have solved the linear programming problems with less than or equal
to (<) or availability constraints using simplex method. In all these problems we
introduced 'slack variables' to convert the constraint inequalities into equalities. Even
if the constraint is 'greater than or equal to type', we converted to less than or equal
to type by multiplying with - 1 through out ( changing the sign). But this method can
be applied if the right hand side of the inequation is already negative.
For example : x^ - x? > -1 can be written as - x^ + x? < 1 , but it is not suitable
for the cases like x± - x9 > 1 which on conversion becomes - x j + x2 < - 1 and troubles
us while calculating minimum ratio ( since - ve min. ratio is ignoivd). Therefore this
method is not suitable in all cases. In such cases we go for one of the following methods.
1. Big - M Method or Penalty Method
2. Two - Phase Method
These are explained here below.

4.3.1 Big-M Method

This is also called Charties Penalty Method.
When all or some of the constraints are 'greater than or equal to' type, then we
choose this method to solve the LPP. The methodology is shown as aflowchart given
below :

Formulate the information

into LPP

Add artificial variable to corresponding

constraints and contribution of
unaffordable cost (-M) for each
artificial variable in objective function
1
| Find the IBFS |
|
Build the table with given
data using the artificial
variables as basic variables
i
Proceed as in
simplex method

FIGURE 4.2 : FLOW CHART OF BIG - M METHOD

Simplex Method 119

Algorithm : For Big-M or Penalty Method :

Step 1: Formulate the given information into an LPP

Step 2 : Convert the inequations to equations by introducing 'SURPLUS' variables

(subtract) followed by adding an 'ARTIFICIAL' variable ( for >) or add slack
(for <) and rewrite the objective function in maximisation form with zero
contribution of each surplus or slack variables but '- M' for each artificial
variable. The ' - M' indicates high penalty.

Step 3 : Find IBFS by keeping artificial variables and slack (if any) in the basis and others
equated to zero (non- basis)

Step 4 : Develop initial simplex tableau -1 and proceed as usually, i.e., find Z, which is
sum of product of Cg value ( - M for artificial variables and zero for slack
variables) and value against corresponding column in the coefficient body
matrix (%•). Then fmd Z - C- to identify key column (most negative value of
Z: - C-) and thus the corresponding 'entering variable'. Find minimum ratio
dividing solution value with key column value to identify key row and thus the
corresponding leaving variable'.

Step 5 : Iterate to develop next simplex tableau.

- Replace leaving variable with entering variable along with its Cg.

- Make key element (intersection element of key column and key row) as
unity, i.e., Divide entire key row of iteration tableau - I by key element
(except it's C$ value i.e., first column), to fmd the corresponding new row
of iteration tableau - II.
- Other elements of key column of iteration tableau - I are to be made as
zero. To make this add or subtract required number of times the entire
key row of Iteration tableau - II in the elements of corresponding rows of
iteration tableau -1

- Find Zj9 Z - Cj and minimum ratio etc., to continue further and repeat
(step 4 & 5) till all the values of Z - C become positive.
Note : In this method when an artificial variables is replaced, it's column also is to be deleted from the
tableau for further iteration.

Step 6: Obtain final iteration tableau and write solution value and hence find the
optimum solution in objective function.
120 Operations Research : Theory and Practice

The above algorithm is illustrated through the following numerical examples.

IllUSTRATiGfy 7 — — -
Food X contains 6 units of Vitamin A per gram and 7 units of Vitamin B per gram
and costs 12paisepergram. Food Ycontains 8 units of Vitamin Aper gram and 12
units of Vitamin B and costs 20 paiseper gram. The daily minimum requirements
of Vitamin A and Vitamin B are 100 units and 120 units respectively. Find the
minimum cost of product mix. Use simplex method.
(JNTU - ECE - 7)

Refer illustration 4 of chapter 2 for its formulation

Summary for Formulation :
Min z= 12x 1 + 2Ox 2
10
S.t. Gxj + S ^ °
7 x j + 12 x 2 > 120
Xj , x2 > 0
Conversion of inequations to equations by introducing surplus variables &
artificial variables.
6xl + 8x2-Sl+Al = 100
7xx + \2x2-S2+A2 = 120 Sx ,S2>0
Revised Objective Function :
Min Z = 12 xx + 20 x2 - OSX - OS2 + MAX + MA2
Max. Z = - 1 2 xx - 20 x2 + OSl + OS2 -MAA -MA2
Initial Basic Feasible Solution :
Basic variables are^4j ,A2 and others are non basic.
,4, = 100 A2=l20 S! = 0 S2 = 0 X1 = 0 X2 = 0
"Incoming Variable
ITERATION TABLEAU -!:

C
J -12 0 0 -M -M
C B.V Min Ratio Remarks
s.v xx s2 A2

100
-M 100 6 -1 0 1 0
8

;®
; 120
:-M\ 120 12 ;o ; o : ;/••• 1

-13 M M M -M -M
( Z
J

r i -13M + 12 - 2 0 M + 2 0
z c M M 0 0

t
Key column
Out going variable
Simplex Method 121

Note : J2 j$ the key element, R indicates new or current row value and R is to be read from previous
table. At becomes non basic and X2 becomes basic variable. Also A% will be removed in the next
iteration.

ITERATION TABLEAU - I I : "Incoming Variable

c
l -20 0 0 -M
C B.V Min. Rat. Remarks
s.v ) •
52

/B
4 9 15
20 0 -1 1
3 3 <- K R

: i_ 1 120
90' 10 1 0 0
12 12 7

.. 4M 140 2M 20
Z
J -20 M + -M
";• 3 ~\ 12 3 12

Out going variable Z}-C} 1 2M 5

4M- K
3
0 M
"3
0

Key column
Note : A\ becomes non basic and x\ becomes basic in next iteration. AndA\ will be removed.

ITERATION TABLEAU • H I :

C
J -12 -20 0 0
B.V Remarks
S.V X-i Xt.y >J 1 J<>

-12 x
i 15 • » -! \

o ' i -I
5
-20 x2
4

h -n -20 I | 2,nax = + 205

» » \ I
z
2max = -205
rcj

Since Z- - Cj > 0 for all variables, the optimisation is reached

The solution is

Max Z - - 12 x 1 5 - 2 0 x - = - 205
4
or Min Z - 12 x 15 + 20 x ~= 180 + 25
4
=205
Optimal solution : Min Z = 205 at xx = 15, x 2 = — , Sx = 0, 5 2 = 0
122 Operations Research : Theory and Practice

I L L U S T R A T I O N 8 (With mixed constraints - with maximising case]

Use Big-M method to solve the following LPP:

Maximise
Subject to

and xpx2>0 [JNTU B.Tech. (Mech) 99]

Solution ;
Conversion of inequalities into equalities
2 Xi + x2 -Si +A• 1 - 2
xj + 3 x2 + S2 = 3
x2 + S3 = 4
Object Function :
Max. Z = 3xl-x2-O ^! + O1S2 + O5 3 -7VL4 1
IBFS : xi = 0, x2 = 0, Si
ITERATION TABLEAU ^ I n coming Variable

C
J 0 0 0 -M Min
cB B.V
s2 ratio
S.V A
i
%r~000r
-•;-;,;:..;4-;.; : ;.:;.:;-; ; v---;l _•_-. ,;
II EiBi
0 3 ^:.;^1^^- 3 0 i 0 0 3
. ; • • . ' . - • • ' • v - . ; • - ' " : • . • ' " • ' • • " '

0 S*. 4 0 0 1 0
- - : T \ . • ' • - • • ' - 1

:«#::-v. 1
z; Af 0 0 -Af
z,.- C/ Si^g|3c - Af + 1 M 0 0 0
t

m^
Out going variable 1 coming Variable
Key column
»ITERATION TABLE I I :

C
J 3 -1 0 Min
CB B.V Remarks
S.V ratio

1 1 -2 RN -^ — R°
3 x
i 1/2 0
(ignore)
ft •.••;:.-/-,:•-..;,•;
.-
• •:•

, - •
• •••

- ; . • •
•

r
•

v
.

'
" '

-
• . -

. : - . , -
•

. .
-

m iiiiiii
RN -> R°
o /
OO
4 0 i i
(ignore)
I Z. 3 3/2 0 A} is removed from tableau
since it is replaced by xx
Out going variable zrcj 0 5/2 ;-:-^2: , 0 0

Key column
Simplex Method 123
ITERATION TABLE - II!:

C
J
3 - 1 0 0 0 Min
B.V
ratio Remarks
s.v X-i Xn Ji On OQ

3 x
\ 3 1 3 0 1 0

0 Si 4 0 5 1 2 0 R$->2R°2

0 4 0 1 0 0 1
Z
J 3 9 0 3 0
ZJ-CJ 0 10 0 3 0

Since Z- - Cy > 0 for all variables, the above solution is optimal.

T h e solution is x^ = 3, x 2 = 0, S\ - 4, 5 2 = 0, 5 3 = 4

Zmax = 3 x 3 - 1 x 0 + 0 x 4 + 0 x 0 + 0 x 4 = 9

Optimal solution : Z max = 9 at ^ = 3; #2 = 0

I L L U S T R A T I O N 9 (With mixed constraints including equality constraint)-

Solve thefollowing using Linear Programming Method.

Maximise Z= -4xx -3x2
Subject to 3xj + x2—3
3xj+4x2 >6

xpx2>0 [JNTUB.Tech.(CSE)97]

Converting the constraint inequations into equations.

Sxi +x 2 +AX = 3 IBFS:

3 X\ + B.V:Al = * 42 = 6 andSj = 4
NBV:Xl = 0, x2 = 0 and 5c, = 0
5] = 4
, x 2 > 0, > 0

Assigning the largest negative price - M to the artificial variables A^ , the

objective function becomes Max. Z-- 4xj - 3x2 + 05 j + 05 2 -MAj -MA% .
124 Operations Research : Theory and Practice
ITERATION TABLEAU I :
' Entering Variable

-3 0 0 -M -M
c B.V
s.v Si s2 Ratio
Remarks

-M 3 \ l . o ;;;..;;o:;; 1 ——Key row

2 6 4 0 -1 0 1 4/3

0 Si 4 i 1 1 0 0 0 6
ZJ-CJ -6M+4 -5M+ 3 0 +M 0 0

Leavin a variable
Key column " Entering Variable
s [ITERATION TABLEAU I I :

-4 0 0 -M Min.
B.V Remarks
Ratio
s.v
-4 1 1 1/3 0 0 0 3

-M ', • • . • • • > • ; . • . ; • •
0 3 0 -1 1 \ 1/3 "«-.,, /X9 ? 2 "^ 1

0 / 3 0 2/3 \ 1 0 0 15/2
(-3M+5) ^4] col. is deleted
0 0 + M 0

1
ITERATION TABLEAU - ! l i :
f
Key column

-4 -3 0 0 Min.
Remarks
x2 52 Ratio
*1 Si
-4 x
l 2/3 1 0 0 1/9 -ve

-3 x2 1 0 1 0 -173 1

0 7/3 0 0 1 + 2/9 43/2 i?3 —> i?3 — — i?2

0 0 0 5/9 A% col. is deleted

As Zy - C; > 0 for all variables, the optimal solution is

X
2 X
l ~ "o" ' 9= 1

Max.Z = -4f|]-3(l) = - y .
Simplex Method 125

Significance of Artificial Variable :

In case of 'Less than or equal to type* inequality, we convert into equality by

adding a 'slack' variable.

Ex : 3xj + 2x2 < 15 is converted as

3xj + 2*9 + 5T = 15 so that 5j > 0 provided xx > 0 andx 2 > 0.

But in 'greater than or equal to type' inequality for instance, as, 5x± + 6 x2 > 9
and xj > 0 x2 > 0 we have to subtract a variable called 'surplus' variable and thus
5xL + 6x2 - S 2 = 9 andx x ,x 2 > 0. Here S2 can not be negative since the negative nature
of 5 2 violates the inequation. Therefore is 5 2 > 0. Now as x t > 0 and 0 x2 > 0, at no
production or 'stand still conditions' i.e., xY = 0 andx 2 = 0, makes the equation
inconsistent with reference to the condition that S2 > 0 (If
x'i = 0 8c x2 = 0 => S2 = - 9 which is violated by S2 > 0). To solve this riddle we have
to add another variable in the equation called 'Artificial variable' (say 'A') so that
A > 0. And A is supposed as very large (the can never be zero) that costs
unaffordably. In Big - M we consider its cost as 'M' a very high cost and concentrate
on its removal while in two-phase method, though we assume unit negative cost,
fundamentally we proceed to remove these artificial variables in phase -1 with out
which we will not enter phase - II. If <Artificial variable' could not be removed in
phase - 1 , we declare there itself that the solution is infeasible.

4.3,2 Two - Phase Method

As the name suggests, we solve LPP in two phases here. The objective function
is split into two sub-objectives, one with artificial variables (and slack if any) only, while
the second with the decision variables. If the phase -1 yields a solution we proceed to
phase - II, otherwise we conclude at phase -1 as infeasible solution.

Advantages of Two Phase Method Over Big-M Method :

The advantages of this method over Big - M method are.

1. It is easier to calculate as it does not involve 7lf and all are numericals.

2. It can give the solution at the first phase itself if the LPP is infeasible. We need
not go through second phase.

3. In th£ case of a digital computer, it is not possible to get the solution by

Big-Af where as 2 - phase method can be applied.
126 Operations Research : Theory and Practice

The Flow Chart for Two Phase Method :

Start

Prepare standard LPP by introducing slack/surplus/

artificial variables write coeff L-V for artificial
variables in objective function

Split objective function into two phases phase-1

with artificial variables phase-II with others

Find IBFS for phase I, identify basic/

Non basic variables and solution values

Set up initial
simplex tableau

Compute Zj - Cj for all columns and

identify most negative as key column

Find key row by

solut ion valu
Min. ratio (,key ,
column value
f )

Iterate to make-key element unity and

other elements of key column as zeros

Is an
artificial variable Solution infeasible
replaceable?

Re-write simplex with objective of phase II

Iterate with usual simplex rules

Find solution

Stop
FIGURE 4.3 : FLOWCHART
Simplex Method 127

The Algorithm of Two-Phase Method :

The following steps are involved in 2 - phase method.

Phase-I:
Step 1: Formulation of Problem : After formulating the information to an LPP with
usual method.

(a) If the objective function is to maximise proceed to (d)

(b) Minimisation objective function is to be converted into maximisation
form. This is done by multiplying with ' - 1 ' through out.

(c) Convert inequality form of constraints to equality form by using artificial,

surplus or slack variables.

(d) Rewrite the objective function (say Zj) by assigning ' - 1 ' value to the
coefficients of Artificial variables (instead of- M as in Big M method) and
'Zero to all the other variables.

Step 2 : Initial Tableau:

(a) Find IBFS, by assigning zero value to decision and surplus variables. Find
the solution values for artificial variables and slack (if any) variables.

(b) Develop initial tableau with usual notations.

Step 3 : Iteration :
(a) Examine the optimality condition by checking if there is any negative
value for Z - C . [Zj is summation of products CB value with corresponding
column coefficient].
(b) Find 'key column' whose Z - C is most negative. Note its corresponding
variable as 'the entering variable'. This replaces the leaving variable in the
next iteration along with its Cj value.
(c) Find 'key row' with minimum ratio by dividing the solution value by key
column value. Note the variable of this row as 'leaving variable'. This,is
replaced by entering variable along with its contribution in the next
iteration.
(d) Iterate with usual method, i.e., to obtain unity at the place of key element
and other elements as zeros in its column, thus find the next tableau.

Step 4 : Basic Feasible Solution (BFS) :

(a) Repeat step - 3 until the values of Z - C- for all the columns are positive.
(b) Obtain the solution called ' Basic feasible solution (BFS)\ This table may
be noted as Auxiliary LPP Solution.
128 Operations Research : Theory and Practice

(c) If Zj - Cj is still negative and Artificial variables are not replaced, stop and
conclude that the LPP has 'infeasible solution*.

(d) Note that when an artificial variable is replaced, it should be deleted from
the table (even its corresponding column also)
Phase - II :

Step 1 ; Resetting the Problem :

(a) Reset the objective function now, without using of artificial variable say
(Z2). Use the original contribution of the variables in 'Maximisation
objective function' for all the other variables.
(b) Use constraint Coefficients as obtained from Auxiliary LPP in phase - 1 ,
i.e., the final (BFS) table of LPP is written as it is, with re-writing the
contributions of all the variables and removing artificial variables.
Step 2 : Initial Tableau :

(a) Obtain initial tableau with the help of the final tableau of phase -1 (Aux.
LPP). Replace the coefficient of variables in objective function with
original values i.e., all C and CB values are re-written according to the
reset objective function.

(b) Check optimality condition and all the process is as same as in step - 3 of
phase -1
Step 3 : Iteration :
Iterate the tableau of step - 2 of phase II obtained above with usual method as
mentioned in step - 3 of phase -1
Step 4 : Optimal Basic Feasible Solution :
Obtain final tableau with all values of Zj - Cj positive (for all columns) and write
the solution value set. Find optimal value by substituting in the objective function of
phase - II.
This is illustrated here below.
I L L U S T R A T I O N 10
Let us consider the example in illustration -1 of Big - M method.
Afterformulation ofthe information, we get
Min Z=12xj+20x2
S.t. 6xj + 8x2>I00
7xj+12x2>120
Xj>0,X2>0 [JNTU B.Tech (ECE) 97]
Simplex Method 129

Now conversion of inequality constraints into equality form by introducing

surplus and artificial variables.
6xx + 8x9 - Sx +AX = 100

xx > 0, x2 > 0, S} > 0 , S2 > 0, Ax > 0, A2 > 0

Revised objective function

Max. Z = - 12*! - 20x2 - 0 ^ - 0S 2 - 1 Ax - \A2
Split the objective function into two sub-objectives
Max. Zx = 0.xx - O.x2 - O.Sj - Q.S2 -Ax - A2
= -Ax-A2 (for phase - I)

and Max. Z 2 = — 12^x ~ 20*2 ~®$l~ ®$2

(for phase - II)
Now for phase - 1 , we have to solve
Max. Zx =-Ax-A2
subject to 6xj - 8 x2 - Sx +AX - 100
7^! + \2x2-S2+A2= 120

IBFS is xx = 0, ^ 2 ^ 0 , 5 ^ 0 , 5 2 = 0 (non - basis )

Ax = 100 a n d ^ 2 = 120 (basis)
ITERATION (INITIAL) TABLEAU - 1 :

(Max. Z | = Oxi + OX9 — 05i + 0So — ^4i —^9)

^ - E n t e r i n g variable

0 0 - 1 - 1 Min
B.V Ratio Remarks
s.v
100
100
- 1 M 8
WSSK&
Ililili
13 &J20;;; 1 1 -1 -1
13 •••-•2KC: 1 1 0 0

Leaving variable
1
Key column
130 Operations Research : Theory and Practice

-Entering variable

Cj 0 0 - 1 Min
B.V Ratio Remarks
s.v
•8 ss
mmmmmimA I?
it
Key element
10 x 12
10 /! 2
• • ^ : \ - • 1 '
12

- 1

Leaving variable

Key column

C
J
0 0 0 0 Min
c. Remarks

is rennoved
B.V Ratio

eau as itis
S.V *\ x2 Si 52

dinba sis
3 1
0 xx 15 1 0
"4 2

0 x2.
5
0 1
7 3
1 ^ ^2 ~*R2--i2R\
4 16 "8
lie
0 0 0 0
z
i-ci 0 0 0 0

As Zj - C > 0 for all the variables and all artificial variables are replaced, phase
-1 yields solution. Therefore we proceed to phase - II
Phase - II :
Objective function Max. Z2 = - 12 xj - 20 x2 - 05 j - 05 2

C
J
-12 -20 0 0 Min
CB B.V Remarks
Ratio
S.V x
\ X
2 ^1 ^2

-12 xx 15
• » -! 1
-20 x2
5^
4 • ' £ -!
Z
J -12 "20 i f

Z
J'CJ • ' i I
Simplex Method 131

As Zj - C: > 0 for all variables, the optimization is reached. And the solution is

5
Xi = 15 Xo = -
1 4
4
Z^ = 12xl5 + 2 0 x |

- 1 8 0 + 25
= 205
ILLUSTRATION 11
Solve thefollowing LLP
Maximise Z=10X+15Y
Subjectto Y>3:X-Y>0;Y<12:
X+ Y < 30; X< 20andX, Y >0 [JNTU (CSE) 2003(Set-1)]
S o I ulion ;

Rewriting the objective function and constraint set in standard form by

introducing slack, surplus and artificial variables, we get,
Maximise Z = 10 X+ 15 K
Subject to

X-Y-S2+A2=Q

X+Y+S4=30
X+S 5 =20
X, Y>0, Sx, S2, 5 3 , 54> S5 >0 and Ax,A2>0

(where X, Fare decision variables, 5j & S2 are surplus variables, 5 3 , 5 4 & S5 are
variables and A { 8c A2 are artificial variables.)

Phase I :
For phase I, we consider the objective function as Max. Zj = -A± -A2
Y-Sl+Al = 3; X-Y-S2+A2 = 0; F+S3=12,

X,Y>0,Sj>0fAl,A2>0 (j= 1,2,.3,4,5)

132 Operations Research : Theory and Practice

IBFS : Basic variables : Ax = 3, A2 = 0, Ss = 12, S4 = 30, S5 = 20

Non basic variables : X = y = S 1 = S 2 = 0

^^~—-Entering variable delete column for next
ITERATION TABLEAU • f : iteration

-1 0 0 0 0 1 - ve (Ignore)

0 1 0 0 0 0 12/1

0 0 1 0 0 0 30/1

0 0 0 1 0 0 oo (ignore)

1 0 0 0 -1 -1

0 0 0 0 0 0

Key column

r
Note : We can choose X or Y as key column here, but better to choose Y and ifX is taken, it may lead to
-—-^Entering variable Delete column

p
TERATIQN TABLEAU

CB B.V ! ° 0 0
s2
0 -1
s4
-1
A2
Min.
Ratio. Remarks

mammnil m
0 Y j 1 -1 0 0

III •HP

1
• • • , - : illll
sglilSll
|

0 9 | 0 1 0 0 0 oo R
o

0 27 ! 0 0 1 0 0 27/1
o

0 20 0 0 0 1 0 20/1
o

Leavirig variable Z
J I 1 1 0 0 0 -1

z c \! o
o 1 1 0 0 0 0
r i
Key column

Note : In the above iteration, we can observe that there is no need of changing R\ and R$ since the desired
1 and 0 are already available
Simplex Method 133
ITERATION TABLEAU HI:
C
J 0 0 0 0 0 0 0 Min.
B.V
Ratio. Remarks
S.V X Y Si s2 s4 $5
p/V . p()
0 Y 3 0 1 -1 0 0 0 0

0 X 3 1 0 -1 -1 0 0 0

0 h 9 0 0 -1 0 1 0 0
0 s4 24 0 0 0 1 0 1 0
0 17 0 0 1 1 0 0 1 Bg^Rl-R"

h 0 0 0 0 0 0 0
z
rcj This is now called auxilary
0 0 0 0 0 0 0
simplex tableau

Since Z - C- = 0 for all the variable phase - I computation is complete at this

stage. Both artificial variables have been replaced from the basis. Therefore we
proceed to phase - II now. Now for phase - II, we take above auxilary simplex tableau
with all numericals 'as is' except for the values of C'•. These are taken from the new
objective function of second phase.

Phase I I :
-Entering variable
Objective function : Max. Z 2 =^

c; 10 15 o/ 0 0 0 0 Min.
C
B B.V
S.V X Y (p) h h s4 s5 Ratio. Remarks

15 Y 3 0 1 -1 0 0 0 0 — ve

10 X 3 1 0 -1 -1 0 0 0 -ve

0 S-i 9 0 0 -1 0 1 0 0 -ve

0 27 0 0 0 1 0 1 0 oo

0 17 0 0 1 1 0 0 1 20/1 <-

f 10 15 -25 -10 0 0 0
z c
I rj 0 0 -25 -10 0 0 0
1
Key column

Leaving variable
134 Operations Research : Theory and Practice

10 15 0 0 0 0 0 Min.
c. B.V C
J V V ^ ^ ^ ^ ? Ratio. Remarks
S.V A I J j "Jg 3 4 r
)

15 Y 20 0 1 0 1 0 0 1

10 X 20 1 0 0 0 0 0 i

0 $3 26 0 0 0 1 1 0 0 r>N v pO , nN

0 $4 24 0 0 0 1 0 1 0

0 17 0 0 1 1 0 0 1

Zj 10 15 0 15 0 0 25 Z m a x = 10x20
+ 15x20
zrc} 0 0 0 15 0 0 25 = 500

Since all the values of Z- - C are positive (> 0), we arrived, at optimal solution.

The optimal solution is X'= 20; Y= 20

Z m a x = 10 x 20 + 15 x 20 = 200 + 300 = 500

Special Cases in Simplex Methods :

1. Infeasible Solutions :

In Big- M method, we find infeasible solution if all the artificial variable do not
leave, the basis.

Similarly, in two - phase method also, if phase -1 does not yield any solution
since all artificial variables can not be replaced, the solution is infeasible.

These are illustrated through the following numerical examples.

Simplex Method 135

ILLUSTRATION 12 -
Max.

5x1 + 6x2>15
x19x2 >0
Solution :
By Big - M method :
Conversion to equality
Max. Z = 4xl + $x2 + 0Sl + 0S

5 x± + 6x2- S2+A= 15
xhx2,Sl,S2,A>0
IBFS xx = 0, x2 = 0, S2 = 0 (Non-basis)
/I-15 (Basis)
- Entering variable
ITERATION TABLEAU !:

C
J 4 0 -M Min
B.V Ratio Remarks
S.V Xi

-M 15 6

-5M .. - 6 M M -M
0
Entering variable
column

C
J 3 0 0 -Af Min
C B.V Remarks
S.V s2 Ratio

0 0
~i ^ row

i 3
-M 1 A 15 0 -1 1 12 R»-*R«-6R»
2 2

j
Leaving variable
Z
J
";• M
:
- 9
4; 3
3 3M
Af -Af
; M " • .7:- 3M 3
z
r
r

c
i 0 + M 0
•"' 2 A
4 T 4
Key column
136 Operations Research : Theory and Practice

^ITERATION TABLEAU -111:

4 3 0 0 -M Min
B.V Remarks
Ratio
S.V x} x2 Sj S2 A

4 x
\ 9
• ! 1 « •
-Af A 5 o -| - f -. .
h 4 f t f f + i M .,,
ZJ-CJ
» ¥+if+i - •
Now all the values of Zj - C.- > 0 but artificial variable did not vanish. Hence
the Solution is Infeasible :

(For the above problem if solution is written it is JC1 = 2, A = 5 , x 2 = 0? S^ = 0

and 5 2 = 0 Z m a x = 8 - 5 M which is not feasible as M is enormously high value that is
unaffordable)

Now , we verify this by 2 - phase method.

By Two - Phase Method :

T h e above problem is done by two - phase method as follows .

For phase - 1 Max. Z^=-A

For phase - 1 Max. Z 2 = 4xj + 3 x 2 +05 - 0S 2

(And the constraints are converted to equality, as same as in the above problem)
Phase - 1 :
'Entering variable
ITERATION TABLEAU i:

CB B.V
4 7./0jM ° ° ~M Min
Remarks
S.V Ratio

ill 1111111 4 •—"^"Key r o w :

"•'••

Z( A 15 5 -;-: 6

\
; • o - 1 i
, . • - • - - . • "6"
-5 ; -6 0 1 - 1
Leaving variable
zrc} -5 %<^t ° 1 °
Key column
Simplex Method 137
ITERATION TABLEAU I I : 'Entering variable

3 0 0 -Af Min
CB B.V Remarks
s2 Katio
x2 s\
1
A
o 0 row •#*&} ,•
Key element
i . •

6 1 0 -1 1 12
; 2 ~ 9^

1 Z
;
v
• • • • . ' : i " • • ' • • •

0
3
9
1 - 1
Leaving variable • • > - . 2 ' ;"

3
- • . - • • • • - • .

0. 1 0
: i
" • " • • / " - i •

Key column
ITERATiOftl TABLEAU !H:

C
J 4 3 0 0 -M
B.V Remarks
S.V Xj X9 Sj 5g A

0 x
\ 2 '. ! 5 » •
- 1 A 5 o -f -| - . pN . DO 1 piV

h • I I • "'
z c
r j • I I ' •
As all the values of Z- - C- > 0 , we stop here. But the artificial variable c^' has
not been replaced. Therefore phase -1 does not yield any solution.

Solution is Infeasible
Unbounded Solution :

When an LPP does not give finite solution values of variables, such solution is
unbounded. In such cases x- can take very high values without violating the constraints
and conditions. This is interpreted in final tableau of simplex as all ratios will be either
negative or infinity so that key row can not be generated and no variable does leave
the basis.

This is already shown under ordinary simplex model and now it is illustrated
with Big - M and 2 - phase methods.
138 Operations Research : Theory and Practice

ILLUSTRATION 13
Maximise Z=3xj+2x2
Subject to x2 -x2<l
Xj+x2>3
x19 x2>0

S>O\L\Y\OV\ I
Solution by Big - M
Conversion :
M a x . Z = 3 xj + 2 * 2 + 0.5-! - O.S2 - MA
S.t. x± - # 9 + 5 ^ = 1
xx +x? - 5 2 + A = 3
xx , x 2 > 0 , Sl>0> 52>0,v4>0
IBFS
xi =0, x 2 = 0, 5 2 = 0 (non basis)
Sx=l .4 = 3 (basis)

ITERATIQSy TABLEAU • 1 :

C
i 0 0 -M Mm
cB B.V Ratio Remarks
sy Si 52 A

HP mm ,:--^,:-4i?R'- •• :•:
^—
] ill l|iig| illiiiiiii
_i i
-M ( A 3 / -. ... . 1 0 12
Z- •":;/" - U : i:
j^ J - ••". -\ - M 0 M -M

Leavin g variable
ZrCj %;3iy - M - 2;
0 M 0
: • • "

Key column

ITERATION TABLEAU - 1 ! ; ^ Entering variable

3 0 -M Min
B.V Remarks
Ratio
s.v 5, A

1 1 0 < •

•2"- ; 0 COluiHll

M Af -M
Z
J-CJ M M
Leaving variable
Key column
Simplex Method 139
ITERATION TABLEAU - HI:
C
J 3 2 0 0 Min.
c. BV
SV v Y c c Ratio
Remarks
x X
l 2 °l °2
x 2 • o I -i
CO

x2 1 o , -1 -i - ve nN pO
CM

2 c) 2

2
Z
J
« 1-I
As Z- - Cj is still negative for at least one of the variables, but the min. ratio is
not possible (- ve for both Xy, x2 variables), we can not proceed further and has no
finite solution.

Solution is Unbounded
Note : Students may try the same problem with two pliase method and check.

ILLUSTBATIG&114 —
Max. Z=4xj+3x2
S.£ Xj-2x2> -10
x2-x2> -12
xpx2> 0
conversion :
Objective Function of Phase -1 : Objective Function of Phase - II :
Max. Z = -Ai -A% Max. 2 = 4xy + 3 x 2

S.t. xl-2x2-Sl+Al=- 10 - O.Sl - O.S2

S.t. Auxiliary LPP Tableau values

Phase -1 :
ITERATION TABLEAU - 1 :
4 3 0 0 -1 - 1
BV Remarks
*i x2. A2

-1 - 10 1 -2 —1 0 1 0 — ve

-1 ^2 -12 1 -1 0 -1 0 1 -ve

h -•2 3 1 1 1 1
zrq -2 3 1 1 0 0
140 Operations Research : Theory and Practice

As Z - C: < 0 for x1? it is entering variable, but no basic (artificial) variable is

ready to leave since min. ratio can not be found (-ve for both rows). We can not proceed
further
Solution is Unbounded

Note : If the above problem is converted into canonical form as

Max. Z=

S.t. - 1 2

- X j +X2< 12

We can do it with out any surplus and artificial variable but simply by
introducing slack variable. But there also we get same problem and the solution is not
finite (unbounded), (students are advised to verify).

Multiple Optimal Solutions :

As studied in the earlier topics, the same condition if seen in simplex tableau,
the solution is multiple optimal. (Refer Sec. 4.3.2)
We know that in an optimal solution, we have
1. Z - C = 0 for all basic variables.
2. Z• - C > 0 for non-basic variables

In addition to the above two points, a multiple optimal solution is found if,
3. Z - Cy = 0 for at least one of the non-basic variables

4. Zopt remains same for all solutions. Usually all the decision variables will not
enter the basis but Z - Cj>0 for all variables here.

ILLUSTRATION 15
Minimise Z=20x3 + 40 x2
Subject to Xj+2x2>3
x2+x2<4

0<*i<-2
0<x2<l_
2
Simplex Method 141

Solution :
Conversion into equalities by introducing slack, surplus and artificial variables
and rewriting the objective function :
Objective function Max. Z = - 20 xx - 40 x2 - 0.5x + 0,S2 + O.S3 + 0S4:-MAl
xi + 2 x^- Sj +Ay = 3
+ X
x
\ 2 + ^2 =
^

* ! + S3 = -

3
x2 + S 4 = -

IBFS: = 0, X2 = 0, 5i = 0 (non basic)

5 3
= 3, S 2 = 4, S 3 = — and S4.= — (basic)

ITERATION TABLEAU I ;

C BV
sv
Y3v; : • • • - . • • " .
- 20

• " • • • • • ; -
V

. . • " • •
If 2
^

.•••-'• • " • • • " .

0

. ' i .
0
s2
0 0

0:
-M Min.
Ratio

9
Remarks

- — - K e y row

Key element
4
0 4 1 .1v 0 1 0 0 0
1

.••-.'•• • '•'•''
Tie

0 5 1 ;;; 0 0 1 0 0
2 neglect
3
0 0 •:/• -.: 1 : 0 0 0 1 0 Key row
• • 1
2
Leaving variable
- M :;;•! :-;:: M 0 0 0 -M
z c
r j - M + 20 - m M 0 0 0 0

Key column

There is a tie for key row in this problem, which is called 'DEGENERACY' in
simplex tableau. In such cases, we choose arbitrarily. However, while deciding the
leaving variable we give preference to artificial variable when tie exists between
artificial, slack and decision variable. [We prefer slack if Tie exists between a slack and
decision and we choose arbitrarily if there is tie between two artificial or two stacks or
two decision variable]. (This is explained in next section clearly).
142 Operations Research : Theory and Practice

ITERATION TABLEAU 2 :

-Entering variable

0 Min.
BV Ratio Remarks
sv s2

-40 I o

0 o
2

ills i^iS0S0MM^^^WM

0
1

r -40 +20

Leaving variable 0 20

Key column (for verifying multiple optimal solution)

From the above table as Z - Cj>0 for all the variables, the optimality is reached.
Thus the solution values are

and Zmax = - 2 0 x 0 - 4 0 x ~

= -60
or =
6® units
Check for Multiple Optimal Solution :

However, the above tableau indicates possibility of getting an alternate optimal

solution. Since, the solution is obtained without replacement by all decision variables
(i.e., X| has not entered basis), and Z - C: = 0 for xy column, this indicates is a scope
for multiple optimal solution. We find this by re-iterating Tableau - II by selecting
x± column as key column and 5 3 as leaving variable.
Simplex Method 143
ITERATION TABLEAU 3:
C
J -20 -40 0 0 0 0
C
B BV Ratio Remarks
SV Si 52
j_ 1 1
-40 X, 0 1 0 0
4 2 2 1 1 o o
5 1 1 pJV jpO /?^
0 s
0 0 1 0
4 2 2 2
,5
-20 1 0 0 0 1 0
2
5 1 1
0 S4 0 0 0 1 /L4 — xi^ + — /13
4 2 2
Z
i -20 -40 +20 0 0 0 2,nax = - 6 0
z c
r i 0 0 20 0 0 0 Z min = 60
From the above tableau - III, we get Z- - C-- > 0 for all the variables, therefore
the optimal solution is reached. .The alternate optimal solution is
5 1

Zmax = - 20 x I - 40 x I = - 50 - 10 = - 60
Z 60
min =

4.4 Degeneracy in Simplex Method

While solving an LPP, the situations may arise (i) in which there is a tie between
two or more basic variables for leaving the basis.i.e., equal minimum ratios or (ii) one
or more basic variables in the "solution values" column become equal to zero. (Such
problems are ignored intentionally in the problems dealt so far in this book, with an
intention to explain here separately). These two cases are called Degeneracies in
simplex problem.
Degeneracy may occur at the initial tableau or at any subsequent iterations.
A degeneracy is detected by the following points.
Case (i) ; Beale's Cycling :
If one are more basic variables contain solution value as 'zero', then the iteration
in simplex tableau repeat (cycle) indefinitely without arriving at an optimal solution.
Suppose a simplex tableau has x1? x2, 5 l s S2 and S3 of which
Sp S2 and 5 3 are in basis. Now suppose xx replaces S± in first
iteration, x% replaces x± in second iteration, Sj replaces x^ in
third iteration and so on. This will continue x1 —> S^ ; x% -» x^
and Si '-> x2 as cycle and can not yield any optimal solution.
144 Operations Research : Theory and Practice

Beale has first detected this type of problem and is explained in illustration 16.
Case (it) : Tie for Leaving Variable From the Basis :

Another interesting case of degeneracy is found with same minimum ratio for
two or more basic variables leaving the basis. This automatically raises the confusion
in selection of key row. In such cases selection may be arbitrary.

This degeneracy can also be resolved by the same method as given for cycling
problem. However, more simplified method to avoid more calculations and minimise
the number of iterations to arrive the optimal solutions are given below. [One example
can be observed in illustration - 15]
Resolution of Degeneracy : (for Cycling Problem)
The degeneracy, when occurs particularly due to cycling, it is resolved by the
following rules.
Rule 1: Divide the coefficients of slack variables (in the simplex tableau where
degeneracy is detected) by corresponding positive numbers of the key column
in the row, starting from left to right.
Rule 2 : The row which contains smallest ratio comparing from left to right column thus
becomes the key row.
Resolution of Degeneracy : (In Case of Tie)

In general the selection of key row is made arbitrary if minimum ratio is same.
(But this selection may increase number of iterations. Therefore use the following
rules).
This are shown as tabular form given below.
S.No. Tie for Leaving Variables Preference to Select as Key Row with
1. Between artificial & slack variable Artificial variable
2. Between artificial & decision variable Artificial variable
3. Between slack 8c decision variable Slack variable
4. Between surplus & decision variable Surplus variable
5. Between surplus & slack variable Surplus variable
6. Between artificial & artificial variable Arbitrary variable
7. Between slack & slack variable Arbitrary variable
8. Between decision & decision variables Arbitrary variable

Thus order of preference is summarised as A > S > X where A is for artificial,

S for slack (or surplus in rare cases) and X for decision variables. (To remember note
that they are in alphabetical order of preference). And arbitrary selection for like
variables such as A 8c A, S 8c S, or X8cX.

The above degeneracies can be observed in the following problems.

Simplex Method 145
ILLUSTRATION -16:

Use degeneracy principles to solve thefollowingLLR

Max. Z=3xj + 9x2
S.t x1 + 4x2<8
x2+2x2 <4
xp x2>0
Solution ;

Adding slacks to convert into equalities, we get

Max. Z = 3 xx + 9 x2 + Q.SY + O.S2

x X
\t 2 — "» ^1» ^2 — 0

IBFS: 5j = 8, 5 2 = 4 (basic)

* l = 0, x 2 = 0 (non-basic)

ITERATION TABLEAU - 1 :
ncoming variable

C
18!!ilS Min..
J
Ratio
5F Remarks
sv KCV\

IJiSlilSl liiiiii^
liili
ljgf§§i
lSIIi

Z;
SMI
2
Tie in out going ,-C; "3

Key column

As there is a tie for out going variable, a degeneracy is indicated. Therefore we

use degeneracy rules by rearranging the tableau - 1.

and $2 a r e brought to left and x^ 8c x2 to right in body matrix)

146 Operations Research : Theory and Practice

REARRANGE TABLEAU - 1 :
-Entering variable

0 0 3 W^^kp Remarks
Ratio
C BV
sv

m
0 8 1 o i cSW9fi^

f
1
Leaving variable 7 ~ ^y
A r
0
o
0
o
0
-3
^:&$i'^
mMf;
Key column

Now applying rules of degeneracy, minimum ratio is found by dividing first

(4 0
column value (FCV i.e., S{) by key column value (KCV i.e., x2), w e § e t among — , —
least is 0, therefore 5 2 is leaving variable. Further iteration is as usual
^•ITERATION TABLEAU II:

C
J 0 0 3 9 Min..
C BV
Ratio
Remarks
SV

0 0 1 -2 -1 0

9 2
2

o f f ,
7 C
» I 4 •
From the above tableau - 2 ,

Since Z - C > 0 for all the variables, we have the optimal solution as follows.

xx - 0, x^ = 2
Simplex Method 147

I L L U S T R A T I O N 17
Afirmmanufacturers two products in three departments. Product fA' contributes
Rs. 5per unit and requires 5 hrs. in dept. M, 5 hrs. in dept. Nand one hour in dept.
R product fB' contributes Rs. 10per unit requires 8 hrs. in dept. M, 3 hrs in dept. N
and 8 hrs in dept. R Capacities for departments M9N,P are 48 hours per week.
Find out optimalproduct mix using Simplex model. Give dual formulation also.
[JNTU-CSE-98]

Solution :
(Refer illustration 3 of chapter 2 for formulation)
S u m m a r y of Formulation :
Max. Z=5XX + 10 X 2
Subject to 5 Xx + 8 X 2 < 48
5 Xj + 3 X2 < 48
Xx + 8 X 2 < 4 8
Xx , X 2 > 0
Conversions to Equations by Adding Slack Variables :

5XX + 3 X2 + S2 = 48
X1 + 8X 2 + S 3 - 4 8
1' 2 *
c c3 > 0
J , kJ2 , p
Initial Solution i

Non basic variables : (X1 = 0 , X 2 = 0)

Basic variables :Sj = 48 , 5 2 = 48 , S 3 = 48
ITERATION TABLEAU 1; -Entering variable

5 0 0 Min..
Ratio
Remarks
S.V xx s2 $s

111
; ;
. - • • : • / • . - . • • • • • • ; - • : ; :

•s.-}. •: ::: . : /t; V .': C Keyrow • ;.:^; ; ; •"•'•;

WImmm
0 s2 48 5 Ill 0 l 0
48
3 Tie

48
0 48 i 0 0 1
8

h 0 0 0 0
Leaving variable
(Selected arbitrarily)
zrc} -5 •••••"^/iO'••"••'.•
• ' • • - . A - . - • • .
0 0 0

Key column
148 Operations Research : Theory and Practice

[The above table has same min. ratio for two rows i.e., degeneracy. As both
variables are slack, we can select arbitrarily or otherwise use degeneracy principle as
. f 8 3 8 , 8, _ f
min. ratio of—, —, — i.e., --, hence choose first row].
M M H ITERATION TABLEAU -1! •:

C
J
5 10 0 0 0 Min..
BV Ratio
Remarks
s.v xx xt Sx S2 S3

10 x2 6
i > i » •
0 s2 30
f » -i • » R$^R%-3R¥

0 0 -1 0 -1 0 1 RX^RO-gRX

h f ,0 f 0 0

f • f » •
As all the values of Z - C- are positive the optimal solution is obtained as
Xx = 0 , X2• = 6; S{ = 0 , S2 - 30 , 5 3 = 0 and

Z - 5(0) + 10 (6) + 0 x 0 + 0 x 30 + 0 x 0 = Rs. 60

The Dual Formulation for the Above Problem : (This is explained in next section 4.6)
Objective function is
Min. Z = 48 Wl + 48 W2 + 48 Wz

Subject to 5 Wx +.5 W2 + Ws > 5

8 ^ + 3 ^ 2 + 8 ^ 3 > 10

Wx, W2 and W$ > 0.

4.4.1 Tie for entering Variable (i.e., Key Column)

In simplex method, some times tie may occur with same most negative value
of Zj - Cj for two or more columns. This tie for key column or entering variable
selection can be made arbitrary. However, the number of iterations may be reduced
by selecting according to the order of preference as decision variable, slack
variable/surplus variable.
Simplex Method 149

.e. • If tie is between decision and surplus / slack variable prefer decision variable.

• If tie is in between two decision variables or two slack/surplus variables choose

arbitrarily.

• There will be no chance of getting a tie with artificial variable since they will be
selected in basis of initial tableau and when they leave basis, these will vanish
from the tableau.

?.g. : Refer to Game theory - Linear Programming method of Solving Rectangular games.

4.4.2 Unrestricted Variables in LPP

Usually an LPP is assumed to have non-negative condition of variables. But in
practical situations it may contain one or more variable with no restrictions i.e., the
variable may be positive or negative or zero. Such variables are known as' unrestricted
variables'.

Since the use of simplex method necessarily requires the conditions of variables
to be non-negative at each iteration, we convert the unrestricted variable into a
restricted non-negative variable by splitting into two parts one as positive part and
the second as negative part. The positive part is any way non-negative while -
negativeof negative part may be taken as again positive i.e., non-negative condition.
Summarily, the unrestricted variable is written as difference of two non-negative
variables.
If %j is unrestricted, let x^-x^ -x^f where xy and Xj" > 0,

4.4.3 Equality Constraints in LPP (or Mixed Constraints)

We have learnt that an LPP, the constraints may be either exact (equality) or
availability (less than or equal to ) or requirement ( greater than or equal to) type.

Since the simplex method demands the constraints to be in equality form for
the computation, use the following rules to make the inequality constraints into
equality type.

1. When the constraint is 'less than or equal to' (<) type, add 'SLACK' variable.

2. When the constraint is grater than or equal to type (>), we subtracted

'SURPLUS' variable. But to satisfy the non-negative condition of surplus
variable we introduce an 'ARTIFICIAL' variable
3. When the constraint is exactly 'equal to' type, it becomes necessary to add-an
'ARTIFICIAL' variable since the exact constraint becomes inconsistent if all
150 Operations Research : Theory and Practice

the decision variables of the constraint assume 'zero* values or non-basic (If at
least one of the variables of exact constraint is chosen in basis with coefficient
as unity i.e., unit vector we need not, add an artificial variable. But it is always,^
safe to add artificial variable for equality constraint).

Summarising we get
Constraint Type Variables Used
< Add slack (+ S)
> Subtract surplus and add artificial (- S +A)
Add artificial (+ A)

The points dealt in the above two topics are illustrated with a numerical
example given below.
ILLUSTRATION 18
Minimise Z=2xj+x2
S.t 3x2+x2=3
4x2+3x2>6
Xj+2x2<4
x2>0, x2 unrestricted.

Solution :

The above problem is converted into equalities by adding artificial, surplus and
slack variables. Also, as x2 is unrestricted, it is re-written as difference of two variables
x 2 ' anck 2 " such that x2 = x2 - x2r where x2 > 0 and x2' > 0

3 x± + x% - x 2 " + Ai = 3

4xx + 3 x{ - 3x 2 " - S%+A2 = 6

Revised Objective Function :

Z = - 2xx - x2' + x{ - O.S2 + O.S 3 - MAX - MA2

where xv> 0, x2' > 0, x2 > 0, S2 > 0, S3 > 0

IBFS : *! = 0, * 2 ' =0 , x2" = 0, S2 = 0 (Non basic)

Al = 3, A2 = 6, S3 = 4 (Basic)
 151

mr
Simplex Method
nina variable
ITERATION TABLEAU -1 :
C
J -1 +1 0 0 -M Min.
cB BV
Ratio Remarks
SV x{ x2" 52 A2

HI
:
3ss>Sft'Kt; r:
-ir:::-.'-::^- r: •'•:•
::
yv^';- &i\

IBi IH tiiiillliS HI;i l l

H i ^ - k - M : .;;••:;/Kft^v*;!'

v : ^ •• •': ; v y A '{•'".;•-•> ' ^ • ; " • • : ; ) " P S

f
fe;tt- • '.z J - ;• • ;'••• •' '• ':-\ '<• :"f itSS.M:-.
%;
•y. >; \tf •:•-:.:•>;:- /'•Wgte&i'iii

6
-Af A2 6 3 -3 - 1 0 0 1
4
4
0 4 2 -2 0 1 0 0
^liiii 1
Out go: ng variable -4M 4M M 0 -M -M i4j is deleted for next
z 4M+ 1 4M- 1 M iteration
r°j
—Incoming variable
Key column

ITERATION TABLEAU - 2 :

A? column is deleted for

ITERATlOm TABLEAU - 3 :
C
J -2 -1 1 0 0 Min.
BV Remarks
SV Xi Xo Xcf Jo Oo Ratio
3 pN . p0 1 piV
-2 x
\ 1 0 0 ^ 0
5 5
6
- 1 0 1 -1 - | 0
5
0 1 r?N . pO 5 pN
0 0 0 1 1 ^ 3 ~ ^ ^ 3 " 3"A2

Z
J
- 2 - 1 1 -j: 0 Zj - Cj > 0 for all variables
5
-••Zmax = - 1 2 / 5
0 0 0 - 0
Z
J~CJ 5 2min-12/5
152 Operations Research : Theory and Practice

Since Z, - C- > for all variables in the above (final) simplex tableau, the optimal
solution is obtained as follows.

Max z
- =- 2 (f)-f=-f orMin - z =f
Note : In the above result it can be further understood that x% always takes a positive value though it is
unrestricted since #2" takes zero value.

45 Concept of Duality in Simplex

One of the most important discoveries in linear programming revealed the
presence of duality that disclosed the fact that an LPP has associated with another
LPP. Thus for the same data we can express two LPP problems, one of which is called
'primal' and the other 'dual'. The relationship between primal and dual is very
intimate and useful that eased in obtaining the solution for the LPP in other way when
one is difficult.
The above concept is explained through an illustration given below.
ILLUSTRATION 19
Prasanna Industries Ltd. manufactures two types of chairs under the trade names
mayurasan andSinhasan. Eeach Mayurasan contributes a profit ofRs, 40 while
that ofSingasan is 50. For manufacturing each Mayurasan, it requires 2 units of
raw material and 8 hrs. of labour while each Sinhasan requires 3 units of raw
materials and4 hrs. of labour. The firm has 30 units of raw material and 45 hrs.
of labour available per day. Formulate t he information as an LPP as primal as
well as dial

Solution :
The above information can be formulated in two ways (say Primal and Dual).
Case (i) : To maximise the profits on the chairs subject to the availability constraints of
the raw material and labour provided the decision is to be taken on the number
of chairs of each type to be produced (decision variables). Obviously the
variables can not assume negative values (production units can not be negative)
and hence we use non-negative condition to the decision variable. Let us say
this problem as 'primal'
Case (ii) : To minimise the utility units or cost (that automatically raises profits)* without
compromising with the minimum profit. Thus the decision variables in this case
will be number of units of raw material to be used and number of hrs. of labour
to the used so that at least the desired profit is obtained. Here also the variable
are non-negative. [Variables need not be non-negative always in all cases. This
depends on the constraints of its primal problem and vice-versa]. This problem
is called 'dual' to the above primal [we can take either way i.e., if case (ii) is
considered as primal then case (i) becomes its dual].
Minimising cost is the other way of maximising, the profits. For example : A
product can be produced in 8 hours that gives a profit ofRs. 100. Now, if you can
make two jobs in 8 hrs. we get Rs. 200 profit and on the other hand if a job can be
completed in 4 hrs. (which gives profit of Rs. 100), it leaves use productive 4 hrs*
worth of producing another Rs. 100.
Simplex Method 153

Let us now formulate the above problems simultaneously in two ways.

Case (i) - Primal Case • (ii) • Dual

No. of units of Mayurasan to be produced = Cost of each unit of raw material to manufacture = w\
No. of units of Sinhasan to be produced = x% Cost of each unit of hrs. of labour to be utilized = w^

Profit on each unit Mayurasan = Rs. 40 Cost on each unit of R/M = w\

No. of units of Mayurasan produced == x\ No. of units of R/M available = 30
Profit on x\ units of Mayurasan = 40xi Cost on 30 units of R/M = 30™ i
Profit on each unit of Sinhasan = Rs. 50 Cost on each unit of labour = w%
No. of units of Sinhasan produced = X9 The labour hours available = 45
Profit on X2 units of Sinhasan = 50*2 Cost on labour = 45z^2
Total profit on both chairs = 40A;i + 50x2 Total cost = 30^i + 45*^2
This profit is to be maximised This cost is to be minimised.
/. Obecjtive function is to .*. Objective function is to
maximise Zp = 40xi + 50x2 minimise Zd = 30u>i + 45^2
SfstSiii
(i) Constraint on availability raw material (i) Constraint on required profits on Mayurasan
Each Mayurasan requires 2 units of R/M R/M required per unit = 2 units
xi units of Mayurasan require 2xi units of of R/M Reqd. profit on R/M per unit Mayurasan = 2un
Each Sinhasan requires 3 units of R/M Labour required per unit = 8 hrs.
X9 units of Sinhasan require 3x2 units ot K/M Reqd. profit on labour per unit Mayurasan = 8^2
Total R/M required = 2xi + 3x2 Total reqd profit/unit Mayurasan = 2wi + 8u>2
This can not exceed 30 available units This reqd profit should produce not less than 40
.-. 2xi + 3x2 < 30 .-. 2wi + 8n/2 > 40
(Constraints on availability of R/M) (Constraint min. profit requirement on Mayurasan)
(ii) Constraint on Labour (ii) Constraint on per unit profit of Sinhasan
In similar way we can obtain the constraint as Similarly we get
8xi + 4x2 < 45 3iv\ + 4u>2 > 50
(constraint on availability of labour) (Constraint min. reqd. profit per unit of

m
Sinhasan)

xi and X2 (the no. of units to be produced) w\,xi>2 (costs of production) are not negative
can not be negative. in this case
.\ x] > 0, x2 > 0 .'. wi>-0, w%> 0
154 Operations Research : Theory and Practice

Summary :

Maximise Zp - 40x\ + 50x2 Minimise Zd - 30wi + 45^2

Subject to Subject to
2xi + 3*2 < 30 • — 2w\ + $u>2 > 40 '

i 8xi + 4*2 ^ 45 ' •— 3^i + 4u»2 ^ 50 '

i t
x\ >0, x2>0 w\ , u>2 ^ 0

Comparison of Solution to the Primal and its D u a l :

ILLUSTRATION 20

Consider the following pair of dual problems: Compare their solutions by

simp*
Primal Dual
Max. Zx = 40x + 50x2 Max. Zw=30w1+45w2
S.t. 2x2+3x2<30 S.t. 2w1 + 8w2>40
3wj+4w2>50
x2>0 w19 w2>0

Primal Dual
Rewriting primal problem in stc I. form Rewriting the dual problem in std. form
Max. Zx - 40xj + 50x9 + OSi Max. Z = - S0w\ - 45w2 -MA\ - MA% - OSi - OS2
S.t. •f 3X2 + S i - 3 U 2 5W2 - il+iil 4U
8xi -I- 4x2 + S2 = 45 B.V. :
l
3w\ + S2 +/12 NBV-X1/X2
xi, x% Si, S2 > 0 W\, Wi §2,A\,i 42 > 0 Si and S2
40 50 0 0 Min. - 3 0 -- 4 5 0 0 -M -M Min.
cB B.V SV C
J
Xi Xo Ratio cB
B.V CJ Ratio
Si SV

0 30 2 3 1 0
30
-M
® 40 2 8 - 1 0 1 0
40

45 50
0 s2 45 8 4 0 1 -M 50 3 4 0 - 1 0 1
4 4
z
r C
i - 40 - 50
T
0 0 M/R +
!2M M M 0 0 M/R
T

50 10 | 1
1
3
0
30
2
-45 5 f 1 -— 0 X 0
40
2

0 s2 5
16
0
4
1 -M A 30 2 0 +~ -1 X 1
30
3 3
20 150 45
z 50 z
r 3 0
3
0 M/R r •c, 8
-2M
0 8 M X 0 M/R

T T
Simplex Method 155
Inversed

50 0 1 -45 0 1 lfi g- X X
2 ~8

40 -30 1 0 X X
4 ~2

0 0
'-> I
2025 Z
J~Ci ° ° ejwx x

Since ail values of Zj - Cj are positive, Sinre, all 7,j~ Cj values are positive, optimal solution for
optimal solution of primal is obtained as dual is rear bed and the solution is w\ = 15, wv = 5/4; and
x] = 15/16^ X2 = 75/8 and shadow prices are shadow prices are 15/16, 75/8
15 and 5/4 Z max = -30 x 15 - 45 x 5/4 = - j 822 %
2025 18225
Z max = 40 x 15/16 + 50 x 75/8 = — ~

Remarks : From above comparison, we can conclude that the solution to a primal problem of linear
programming can always, provide a solution to its dual, check the relationship described
by arrows carefully to understand to comparison.

4.5.1 Conversion of Primal to Dual or Dual to Primal

The following steps (rules) will enable the conversion of any primal problem
into its dual and vice-versa.

Step! : (a) If in the given problem, objective function is m maximisation form

proceed to step 2.
(b) Convert the objective function of the given problem to maximisation
form if it is minimisation form. This conversion can be done by
multiplying with - 1

Ex: Min. Z — x^ + 9 x2 + x% converts to Max. Z = - 2 xj - 9 x%-

156 Operations Research : Theory and Practice

Step 2 : (a) If all constraints have "less than or equal to" (<) sign, go to step - 3

(b) If a constraint has "greater than or equal to" (>) sign, convert it to "less
than or equal to "

Ex •: (i)xl + 4x2 + 2x3 > 5 is rewritten as - x - 4x2 - 2x3 < - 5

(ii) 3xY + x2 - 2 # 3 > 4 is rewritten as - 3xj - x2 + 2xs < - 4

(c) If a constraint has an "equality" sign (=), split this into two constraints in
two opposite inequalities.
Ex: 2xY + 3 x2 = 5 is split into 2 xj + 3 x2 < 5 and 2 xj + 3 x2 > 5
Then these are re-written using step - 2 (b)
2*2 + 3x2 ^ 5
and - 2 x ! - 3 x 2 < - 5

Ste/? 3 : (a) If all the variables are non-negative i.e., xy > 0, then proceed to step 4

(b) If any variable is unrestricted replace it by difference of two non-negative

variables (These variables are to be replaced in the entire problem)
Ex: Ifxj is unrestricted, change it as xj > 0 , xf > 0 where

Step 4 : Now we have standard problem in primal. To write its dual we use the following
principles.
(a) Take a different set of decision variables. The number of variables in dual
will be equal to the number of equations in primal (and vice versa).
(b) We have to minimise the objective function in dual (primal in
maximisation from)
(c) The objective function is written with the constants of inequality
constraints as coefficients of decision variables.
(d) Consider a matrix of coefficients of constraint inequations of primal.
Transpose the rows and columns of coefficients matrix of primal to write
constraint inequations of dual.
(e) Write > sign inequality for dual constraints.
(f) Constants of inequality of dual are drawn from coefficients of variables in
objective function of primal.
Step 5 : Conditions of variables are unaltered i.e.,non negative
Step 6 : Rearrange the dual formulation if required. For example if a constraint has
negative sign on its right side (i.e., for constant), multiply entire inequality with
(-1) and convert the constraint into '<' type. Similarly, re-arrange the
conditions of variables if required.
Simplex Method 157

Primal Dual
Variables
Say, Xj, x 2 , x 3 (3 variables) t^j, tt»2 (2 variables)

J
Objective function :

Minimise: HI, TO,

Maximise: [n^ n^ n3"|
L -J L n '2j

' Constraints
] i n
Wi
0,-y Cic) Q>£
x9 '< h
2
~ ]i
t
*3 r. H h
1 Transposed

(2 constraints) (3 constraints)
Conditions
•x x x > 0 wvu>2>0

ILLUSTRATION 21
Write the dual of the following primal LPP.
Maximise Zp = 50xj + 60x2
Subject to 2x1 +x2 < 300
3xt+4x2<509
4x2 + 7x2<812
xpx2>0
SoLvHcm :
Step 1 . Objective function is max. form therefore proceed to step - 2
All constraints are < type so, proceed to step - 3,
Step 3 . Condition of variables are all non-negative, hence go to step 4
4: (a) Number of variables should be 'three' since there are three constraints in
primal (say w^, tt/2, tu§)
(b) Objective function is to be minimised (since it is in max. form in primal)
(c) Coeff. of variables in dual objective function are taken from constant of
primal constraints viz. 300, 509 and 812
/. objective function is Min. ZD = 300 w± + 509 w^ + 812 w$
(d) Coefficients of constraints set in primal are transposed in dual constraint
set.
3
Thus transpose of is
4
158 Operations Research : Theory and Practice

(e) Constraints in dual take *>' form since primal has all '<' constraints

(f) Dual constraints are taken from coefficients of objective function of

primal, viz. 50 and 60

.-. constraint set is

2wl + 3w2 + 4w3> 50
Wy + 4 w2 + 7 w§ > 60

Step 5 : All the dual variables take non-negative condition.

i.e., w-y > 0 , w2 > 0 and w% > 0
Summary :
iViiwa/ Dual
Max Zp=50xj + 60^9 Min. Z^ = 300 w l + 509w2 + 8l2w3
S.t. 2 xj + JC9 < 3 0 0 S.t. 2 wj + 3 u,2 + 4w3>50
3x 1 +4x 2 <509 Wy + 4 uJ2 + 7 W 3 > 6 0

4^ 1 + 7x 2 <812 w
x1;x2>0

TRATIflW ?7
I n n 1 1 U IV £. £. ~———————————————————————————

Write dual for

Minimise Z=2x1 + 9x2 +x^
5.t x1+4x1+2x3>5
3x2+x2+2x3>4
xpx2,x3>0 [JNTU(Mech)2001]

Solution :
Step 1 : Convert Minimising objective function to maximasation form
i.e., Max Z* = - 2x^ — 9x2 - x%
Step 2 : Convert '>' inequality to < type
- Xy - 4x2 - 2 X3 < - 5 •
~3xl-x2-2x3<-4
Step 3 : All variables are non-negative
xl9X2,x$> 0
Step 4 : For dual,
(a) Variables must be two only (since primal has two constraints) say W\,
(b) Objective function is to minimise
(c) Coefficients are constants of constraints in primal
Simplex Method 159

i.e., - 5 and-4
objective function is Min. ZD - - 5 w x - 4 w2
or Max. Zp =bwy+ Aw^
(d) Coefficients of primal constrains are transposed.
•-1 -3"
. [-1 -4 -2l. ,
i.e., „ is transposed as -4 -1
\-5 -1 -^J
-2 -2
(e) All take > form
(f) Constraints are taken from coefficient of variables in objective function
i.e., - 2 , - 9 and - 1
-2 or
— 4 XVy —7^2 - ~ 9 4wx

(g) Both variables take non negative conditions

i.e., Wp w2>0

Summary :

Standard Primal Dual to Standard Primal

Max. Zp~~ Min. ZD-~ bwx - 4a»
S.t. - Xj - 4 < -5 S.t. -a^ ~3w 2 >-2
- 4wj - u»2 ~ " 9
>0

ILLUSTRATIOW 23 — — —
Write the dual ofthefollowing primal
Min. Z=4xj+3x2
S.t. 3X2+X2=3
3x2+4x2>6
Xj+x2<4
x2+x2>0 [JNTU-CSE-1997]

Solution :
Step 1: Converting objective function min. to max. form
Max. Zp = - 4 x1 - 3 x2
160 Operations Research : Theory and Practice

Step 2 : Written all constraints in '<' form - 3JC1 - x2 < - 3 ; 3x^ + x2 < 3

- 3 xx - 4 x2 < - 6 ; xj + x2 < 4
Step 3 : Condition of all variables are in non-negative form, henc*e proceed
Step 4 : Now, for dual we have,
(a) 4 variables say wyt w{',w2 and w% (since there are 4 constraint in primal
ref, step 2)
(b) Objective function is to minimise, since primal is to maximse (ref. step -1)
(c) Coefficient of obj. function of dual are - 3, + 3, - 6 and 4 (the constants
in constraints of primal)
.-. Objective function is
Min. ZD = - 3 wi + 3 w\" ~ 6 w2 + 4 w%
Min. ZD = - 3 {wi - Wi") - 6 w2 + 4 w§
or Max. ZD = + 3 (w^ - w^') + 6 w2 - 4 w%
(d). Constraint set transposing
•- 3 ~ r
+3 + 1 f - 3 + 3 - 3 1
tO
-3 - 4 [-1- + 1 - 4 1
1 1
(e) Constants are - 4 and - 3 (Since these are coefficients in primal objective
function)

.*. Constraint set is

- 3 w 1 ' + 3a; 1 "--3w 2 + ry3 > - 4
i.e., - 3 (wy - w^') - 3 w2 + w§ > - 4
and — u^j' + Wyf — 4 t^ 2 + tv^ > — 3

or 3 (wi - wx") + 3t^2 - u>$ < 4

(f) All variables are non-negative

Assuming w{ - w{' = w^ and hence w^ unrestricted

We get the dual as
Max,. ZD = 3 w^ + 6 w2 — 4 w%
S.t. 3 wx + 3 w2 - w$ < 4
t^l + 4 w2 - ^3 ^ 3
Wj unrestricted, w2, w§ > 0
Simplex Method 161

Summary :
Standard form of Dual
Kll§lliI Primal (mcj , x2) &{', wv w3)

Max. Zp = - 4xj - 2>x2 Max. ZD = - 3 ^ '

S.t. - 3 * ! - x 2 < - 3 S.t. - 3 ^ ' + 3^2"-

3*2 + x 2 < 3 + w$>-4

-3x1-4x2<-6 ~ W\' + "IVy* — 4Wo

" Xj + JC2 < 4 + M)3 > - 3

x,, x2 > 0 ai/, t t ' i " , Z£J2, ^ ' ^ - ^

Remarks: 1. In the above final dual the objective function can also be Min.
Z-Q — 2>w j - 6 W2 + 4 Wo, or Max. Z^ - - 'Sw^ + 6^2 — 4 w$ since Wy is unrestricted.
2. For a constraint of equality form in primal, we get correspojiding variable unrestricted
in dual.
3. For an unrestricted variable in privial, we get corresponding constraint as equality in its
dual.

ILLUSTRATION 24

Fm^/ the dual of

Max. Zp=3xj +x2 +2x3- x4
S.t 2xj -x2+3x3+x4 = l
x2+x2 -x3+x4=3
xpxPx3>0,x4 unrestricted [Dr. BRAOU (MBA)-Assignment 98]
Solution :
With usual steps, writing the standard form of primal we get
Max. Zp = - x 4 ' + x4

S.t 2^2 - x 2 + 3*3 + x4 - x 4 " < 1 L equality constraint is re-arranged

- 2xy + x2 - 4 1
Xj + x% - *3 ' - x 4 " ^ 3 I equality constraint is re-arranged
- xj - # 4 + X4" ^ - 3
Xp x 2 , x 3 , x 4 ', x 4 " > 0 where x 4 ' - x 4 " = x 4 (since x 4 is unrestricted)
Dual form is
M i n . Zj) = 101—11)]"+$ W2 — $w^r
S.t. %w.{ - %w{f + w{ - w{'> 3
162 Operations Research : Theory and Practice

- w 2 / + w2" >2
Wy -

w
\
Re-arranging with assumptions as it^ = a^7 - o^77 and w2 = ^2' ~ ^2" DY which
and ^2 wiM become unrestricted we get the dual revised as,
Min. ZD = wj + 3 This can be further reduced to
S.t. 2ty 1 + Min. -Zj) = wi.+ 3t^2
St. 2wl+w2 >3
- tfj + t^2 -
>2 - WY + a^2 ^ 1

^ l + ^2 ~ ~ 1 (in!^11 of two inequations)

aj j and t^2 unrestricted.
and w^ unrestricted

45.2 Comparison Between Primal & Dual

1. With Reference to Formulation :
Primal Dual

If objective function is in Maximisation form Objective function is in Minimisation form

If objective function is in Minimisation form Objective function is in Maximisation form

>

Coefficients of decision variables in objective function Constants of constraints set

Constants in constraint set Coefficients of decision variables in objective function

>

Number of decision variables Number of constraints

>

Number of constraints Number of decision variables

Coefficients of variables in constraints row wise Coefficients of variables in constraints columnwise

>

Coefficients of variables in constraints column wise Coefficients of variables row wise

<, type constraints £ type constraints

>

10. a type constraints £ type constraint

>

Equalii
Equality type (=) constraint Corresponding variable unrestricted
>

Unrestricted variable Corresponding constraint is equality (=) type

>
Simplex Method 163
2. With Reference to Solution :
S.No. Primal Dual
1. Solution is taken from solution values Solution is taken Zj - Cj row of final
column of final primal tableau (simplex) simplex tableau of primal
2. Unbounded solution Infeasible solution
3. Infeasible solution Unbounded solutions
4. Solution values Shadow prices
5. Shadow prices Solution values

4.5.3 Advantages & Applications of Duality

Duality will be more advantageous in the following cases.
1. Sometimes dual problem solution may be easier than primal solution
particularly, when number of decision variables is considerably less than
slack/surplus variables.

2. In the areas like economics, it is highly helpful in obtaining future decision in

the activities being programmed.
3. In physics, it is used in parallel circuit & series circuit theory.

4. In game theory, dual is employed by Column player who wishes to minimise

his maximum loss while his opponent while Row player applies primal to
maximise his minimum gains. However, if one problem is solved, the solution
for other also can be obtained from the simplex tableau (Refer game theory :
LPP method)

5. When a problem does not yield any solution in primal, it can be verified with
dual.
6. Economic interpretations can be made and shadow prices can be determined
enabling the managers to take further decisions.

4.5.4 Shadow Price

The shadow price of a resource is the unit price that k equal to increase in profit to be
realised by one additional unit of the resource, (or)
// is the change in the optimum value of the objective function per unit increase of the
resource.
The shadow price can be determined from final simplex tableau using the
values of Z - Cj, in primal problem.

In the illustration - 21 of this chapter, we conclude that

164 Operations Research : Theory and Practice

1. For Primal : A unit increase of resource covered by first constraint raises the
profit of x{ by 15 units and a unit increase in resource covered by second
constraint raises profit of x2 by 5/4 units. Thus shadow prices of xj and x2 are
15 and 5/4 units separately.

2. For Dual : A unit slash in requirement of-first, .constraint reduces the cost of
Wy by 5/16 units while a unit slash on second constraint requirements reduces
cost of w2 by 75/8 units. Thus shadow prices of w.y and w2 are 15/16 and 75/8
respectively.

Practice Problems
Solve the following LP problems by two-phase method.

1- Max. Z = 3 %Y ~x2

Subject to the constraints

2 xx + x2 > 2 ; xj + 3 x2 < 2
4
*2- ' xx,x2>0 [OU- MBA May 92]
Answer: Xj = 2 ; x 2 ~ 0; Max. Z = 6.
2. Max.. Z = 2 x 1

Subject t o : 3xl+ \0x2 + 5x3 < 15;

33 *! - 10 x 2 + 9 x 3 < 33 ; x 1? *-2 , ^ 3 > 0

; Infeasible solution since artificial variables can not be removed.

3. Max. Z = 5 xy + 8 x2 Subject to 3xx + 2 x2> 3 ;

x± + 4x2 > 4 ; Xj+X2 ^ 5 ; xj,X2>0

Answer: x± = 0; x 2 = 5 and Max. Z = 40

4. Max. Z = 5 x^- 2x2 + 3 x$

Subject to : 2 x} +2x2-x$ >2 ; 3 x1 + 4 x 2 < 3

x2 + 3 x3 < 5 ; #i,-#2» ^3 - 0 '
23
= 5 ; x 3 = 0 ; Max.. Z - 85/3

5. Max. Z = ac1 + 1.5x 2 + 2x s + 5 x 4 with the constraints

3 xi + 2 x 2 + 4 X3 + x 4 < 6 ; 2 #i + x 2 •+ #3.+ 5 ^ < 4 ;

2 xj + 6 x 2 - 8 x 3 + 4 x 4 = 0 ; xj + 3 x 2 - 4 x 3 + 3 x 4 =0
* ; - ( / = 1 , 2 , 3, 4) > 0

Answer: xj = 1.2 ; x 2 = 0; x 3 = 0.9, x 4 = 0, Max. Z = 19.8

Simplex Method 165

6. Max. Z = 3 xx + 2 x 2 + x 3 + 4 x 4

Subject to 4 xj + 5 x 2 + x 3 - 3 x 4 - 5 ; 2 xj - 3x,2 - 4 x 3 + 5 x 4 '= 7

=
Xj + 4x 2 + 2.5 x 3 - 4x 4 ^ > Xj , x 2 , X3 > 0

; Infeasible solution

7. Minimise Z = x 1 - 2 # 2 ~

Subject to - 2 xj + x 2 3 x3 = 2 ;

Xj>Of j = l , 2 , 3,

Answer : Here all Z- - C > 0, .but artificial variable A j does not vanish from
the basis. Hence the given LPP has infeasible solution.
8. Solve the following problems by simplex method adding artificial variables.
Max. Z = 2 xi + 5 x2 + 7 x3
Subject to 3xj + 2x 9 + 4x 3 < 100; xx+4x9 + 2x^ < 100

50 50
Answer ; x\ = 0 ; x 2 = — > S X =
~T~» M a x ^ =

9. Max. Z ~ 4 x 2 + x 2 + 4 x 3 + 5 x 4

Subject to 4 xj + 6 x 2 - 5 x3 + 4 x 4 > ™ 20 ; 3 xj - 2x2 + 4 x3 + x 4 < 10;

8 x^ - 3 x 2 - 3 X3 + 2 X4 < 20 ; x^ x2, x3, x 4 > 0
Answer: Unbounded solution
10. Hema & Co. produces three types of pens of types namely sharp, fighter and
pride. These products require three qualities of inks In the refills say
Pi P 2 and P 3 . The quantities of these refills available are 22, 14 8c 14 hundreds
respectively. Each of these pens, the refill requirements are
Sharp Fighter Pride
(S) (F) (P)
3
CO

2 3
t—(

2 0
CO

P3
Profit hundreds 1 4 5
(Rs. in thousands)
166 Operations Research : Theory and Practice

The company makes a profit of one, four 8c five hundreds on each unit of the
products (pens) respectively. How many pens of each type should the company
produce to maximize the profits?

Hint; Formulation is Max.. Z = x^ + 4 x2 + 5 x 3 ;

Subject to 3 * ! +x 3 < 22

Xj + 2 ^2 + ^ X3 ^ 1 4
3*.!+ 2*2 < 14

1.1. A manufacturer of steel furniture makes three produces chairs, filing cabinets
and tables. Three machines (Say P, Q and R) are available on which these
products are processed. The manufacturer has 100 hours per week available
on each of three machines. The time required by each of three products on
three machines is summarised in the following table.

Time Required (in hrs)

Product
Machine P Machine Q Machine R
Chair 2 1
CM

Filing cabinet 1 2
CM

Table 1 2
1

The profit analysis shows that the net profit on each chair, filing cabinet and
table is Rs. 22, Rs. 30, and Rs. 25 respectively. What should be the weekly
production of these products so that the manufacturer's total profit per week
is maximised.

Answer : —— chairs, — tables, — file cabinets. Max. profit = Rs. 1650.

12. A factory is engaged in manufacturing three products X, Y, and Z which involve

lathe work, grinding and assembling. The cutting, grinding and assembling
time required for one unit of X are 2, 1 and 1 hours respectively. Similarly they
are 3, 2 and 3 hours for one unit of Y and 1 •, 3 and 1 hours for one unit of Z
and the profits on X, Y and Z are Rs. 2 , Rs. 2 and Rs. 4 per unit respectively.
Assuming that there are available 300 hrs of lathe time, 300 hours of grinding
time & 240 hours of assembly time, how many units of each products should
be produced to maximise profits? Work the problem using simplex methods.
Simplex Method 167

13. A transistor radio company manufacturers four models A, B, C and/) which

have profit contributions of Rs. 8, Rs. 15, and Rs. 25 on^4, B, and C respectively
and a loss of Re. 1 on model D per dozen. A dozen units of model A require
one hour of manufacturing, two hours for assembly and one hour for packing.
The corresponding figures for a dozen units of model B are 2, 1 and 2 and for
a dozen units of C are 3, 5 and 1, while a dozen units of D require 1 hour of
packing only. During the fourth coming week, the company will be able to make
available 15 hours of manufacturing, 20 hours of assembling and 10 hours of
packing time. Obtain the optimal production schedule for the company.

Hint: Formulation is
Max. Z = 8 xx + 15 x2 + 25 x3 - x4
Subject to x1. + 2-*2-+3*3 ^ 1 5 ; 2 xx +x 2 + 5x 3 < 20
v
Jlj
-L_ f j v
T A AO
_L *V
T Aft
_L *v*
i J^A
<^* 1 ( l
— ±\J
•
y
f\fl
Aj
*\A
j ^Qy
v
^l. >0

Answer: xx=x2 = x$ = — ; x4 = 0 ; Max Z = Rs. 120

14. A manufacturer has two products P 8c Q both of them are produced in two steps
by machines Mj 8c M2. The process times per hundred for the products on the
machines are :
Contribution
(per hundred)
p 4 5 10

0. 5 2 5
Available (hrs) 100 80

The manufacturer is in a market upswing and can sell as much as he can

produce of both products. Formulate and determine optimum product mix
using simplex method.

Hint: Formulation of the problem is

Max. Z = 10 xj + 5 x2 ,

Subject to 4 * ! + 5*2 £-100

5 xx + 2 x2 < 80
#j, %2 > 0

2000
Answer: Xj = 1177; « 2 = - 1059.
17
168 Operations Research : Theory and Practice

15. A manufacturer of leather belts makes three types of belts A, B and C which are
processed on three machines My, M2 8c M$. Belty4 requires 2 hours on M t 3
hours on M 3 . Belt B requires 3 hours on My , 2 hours on machine M2 and 2
hours on machine M 3 . Belt C requires 5 hours on M2 and 4 hours on Ms. There
are 8 hours of time per day available on machine My , 10 hours per day available
on machine M 2 and 15 hours of time per day on M 3 . The profit gained from
Belt A is Rs. 3 per unit, from belt B is Rs. 5 per unit, from belt C is Rs. 4 per
unit, what should be the daily production of each type of belts so that the profit
is maximum.
16. An animal feed company must produce 200 kgs of a mixture consisting of
ingredient Xy 8c X2 daily. Xy cost Rs. 3 per kg and X2 costs Rs. 8 per kg. No
more than 80 kgs of Xj can be used and at least 60 kgs of X2 must be used. Find
how much of the each ingredient should be used if the company wants to
minimize cost? [0U - MBA - Apr. 98]
Hint: Formulation is Min. 3 xj + 8 x2
Subject to xy + x2 = 200; xx <80
x2> 60 ; j , X2 > 0

Answer: xx = 80 ; x2 = 120 ; Min. cost = Rs. 1200

17. A television company has three major departments for manufacturing of its
models supreme and delux. Monthly capacities are given as follows
Per Unit Time Requirements (Hrs) Hours Available
Department
Model Supreme Model Delux Per Month
I 4.0 2.0 1,600
II 2.5 1.0 1,200
III 4.5 1.5 1,600

The marginal profit of model supreme is Rs. 400 each and that of model delux
is Rs. 100 each. Assuming that the company can sell any quantity of either
products due to fovourable market conditions; Determine the optimum output
for both the models, the highest possible profit for this month and slack time
in the three departments.
Hint: Formulation is : Max. Z = 400 xy + 100 x2 ,
5
Subject to 4 xx + 2 x2 < 1600 ; -xl+x2 < 1200

9 3
•^xy + — x2 < 1600 ; yyx2 >0

3200
n n ™ 1280000
7
Answer : Xy = —-— ; x2 = 0 ; x% = 0 ; Max. Z =
Simplex Method 169

18. A teacher gives his students three long lists of problems with the instructions
to submit not more than 100 of them correctly solve, for credit. The problem
in the first list are of 5 points each, in second 4 points each and in third 6 points
each. On an average 3 minutes are required to solve a problems from first list,
2 min. for a problem from second & 4 min. for a problem from tfiird. The
students devote more than 2V2 hours of numerical hours. How many problems
from each list, a student should solve so as to get the maximum credit.

Answer : Max Z = 5x± + 4x% + 6x3

Subj to x + x2 + x 3 < 100; + 2x2 + 4xs < 150

xj , x2 , x3 > 0

19. A furniture company manufactures four models of desks. Each desk is first
constructed in the carpentry shop and is next sent to the finishing shop where
it is varnished, waxed & polished. The number of man hours of labour required
in each shop is as follows :

Desk
Shop
I II III IV

Carpentry 4 9 7 10
Finishing 1 1 3 40
Profit per item (Rs.) 12 20 18 40

Because of limitation in capacity of the plant, not more than 6,000 man hours
can be expected in the carpentry shop and 4,000 in the finishing shop in.a
month. Assuming that raw materials are available in adequate supply and all
desks produced can be sold, determine the quantities of each type of desk to
be made for maximum profit of the company.

Hint: Formulation of the problem is :

Max. Z= 12* x + 2 0 * 2 + 1 8 * 3 + 4 0 * 4 ,
Subject to 4 ^ + 9*2 + 7*3+10*4 < 6000 ;
*! + x2 + 3 x3 + 40 x4 < 4000 ;
* 1 , X2, *g, ^4 > 0

400 20
A ° n rt ° A ™ 7 » 56QQQ
Answer : x j = —-—, x% = 0, * 3 = 0, x 4 = —— and Max. Z = Rs — - —
O O U
170 Operations Research : Theory and Practice

20. A manufacture produces three products A, B and C. Each product can be

produced on either one of two machines, I & II. The time required to produce
1 unit of each product on a machine is given in the table below.
Time to Produce Limit (Hrs)
Product
Machine -1 Machine - II
A. 0.5 0.6
B 0.7 0.8
C 0.9 1.05
. There are 85 hours available on each machine ; the operating cost is Rs. 5 per
hour for M/C I and Rs. 4 per hour for M/C - II and the product requirements
are at least 90 units of A, at least 80 units of B and at least 60 units of C. The
manufacturer wishes to meet the requirements at minimum cost.
Solve the given linear programming problem by simplex method.
Answer: 150 hrs on m/c II, no time on m/c I and min. cost = Rs. 600.

Additional Problems
1. The post master of a local post office wishes to hire extra helpers during the
Deepawali seasons, because of a large increase in the volume of mail handling
8c delivery. Because of the limited office space and budgetary condition, the
number of temporary helpers must not exceed 10. According to the past
experience, men can handle 300 letters and 80 packages per day, on the
average and women can handle 400 letters and 50 packages per day. The post
master believes that the daily volume of extra mail and packages will be no less
than 3,400 and 680 respectively. A man receives Rs. 25 a day and a woman
receives Rs. 22 a day. How many men and women helpers should be hired to
keep the pay-roll at a minimum?
Answer: Min Z = 2 5 ^ + 22x2
S.t. xx + x2 < 10 ; 300*! + 400x2 > 3400;
80xj + 50x2 > 680 and x h x2>0
2. Two products X and Y are processed on three machines My, M2 and M 3 . The
processing times per units, machine availability and profit per unit are as under.
Processing Time (Hrs)
Machine Availability (Hrs)
X Y
M{ 2 3 1500
M2 3 2 1500
M3 1 1 1000
Profit/unit 10 12
Formulate the mathematical model, solve it by simplex method 8c also find the
number of hours machine M 3 remains unutilized.
Simplex Method 171

3. A manufacturing firm has discontinued production of a certain unprofitable

product line. This created considerable excess production capacity and
management is considering to devote this excess capacity to one or more of
three products : X, YandZ. The available capacity on the machines which
might limit output is summarized in the table.

The no. of machine hours required for the unit of the respective product is
given below.

Productivity (in machine hrs per unit) Available time

Machine Type
Product X Product Y Product Z in M/C hrs/week

Milling machine 8 2 3 250

Lathe 4 3 150
Grinding 9
L 1 50

The unit profit would be Rs. 20, Rs. 6 and Rs. 8 respectively for products
X, Y and Z. Find how much of the each product the firm should produce in
order to maximise profit?

Hint: Formulation is : Max. Z = 20 xj + 6 x 2 + 8 x§

Subject to 8 xY + 2 x2 + 3 xs < 250

4xj + 3 * 2 < 150;

2 * i + x 3 <50;

Xi, x2, x 3 >0

Answer: xy = 0; x% = 50, x 3 = 50 ; max. Z = 700

4. A plant is engaged on the production of two products which are processed

through three departments, the number of hours required to finish each is
indicated in the table below :

Product Max.. Hours

Machine Type Available Per week
A B
I 7 8 1600
II •8 12 1600
III 15 16 1600
(a) If the profit for the products is Rs. 6 for a unit of product A but only Rs.
4 for a unit of product B, what quantities per week should be planned to
maximize profit.
(b) Capacity can be increased on one department only, in which department
should it be done 8c why? To what extent should be capacity be increased?
172 Operations Research : Theory and Practice

(c) If the cost per hour in department I is Rs. 25, in dept II, Rs. 40, & in dept
III> Rs. 50; what quantities should be planned to minimize to cost of
production?
320
Answer ; (a) —^— units of A & no units of B ; max. profit = Rs. 640
3

(c) — units of production^ 8c No. unit of B, min. cost = Rs. 20

3
5. Max. 4*+ 10 ^y
S.t. 2x+)><50,
)< 100,

x,y>0
Solve graphically
If solved by simplex method, is the solution unique ? Why/why not? if not, give
two basic optimal solutions. Also find all the non-basic optimal solutions to the
problem.
6. Minimise 2 0 0 ^ + 350*c2
S.t. ••50*i+ 60*2 ^2500
100^ + 60x2 > 3000
100*!+ 200*2 >7000
*!, x2 > 0 [ou - MBA - May 91, Nov. 94, Oct. 95]
7. Mm. Z = 2xx +x2-x>i'~x4
S.t. X!-x 2 + 2
2*1 +%9~
- %Xa + YA = 6
*1 + X2 + x 3 + x 4 7
and > 0 [OU - MBA - May. 92]

8. Max. z =9*i + 15x2 4- 20 x 3

S.t. 3xi +1 < 210
3* 1 + 5x2 + 4x3 < 240
8 x 1 + l C )x 2 + 12x3 < 260
o [0U - MBA- May 9 1 , Nov. 94]

9. Max. z =5*1 + 2* 2 + 10 x 3
S.t. < 10
X
2 ~ XS > 10
+ x 2 +x 3 < 10
*1,.*2,*3 > 0 [0U - MBA May 95]
Simplex Method 173

10. Min.
S.t. 36^! + 6x2 > 108
3xi + 12x2 > 36
20xj + 10x2 > 100

[0U - MBA Feb. 93]

11. Max. Z = SxY

S.t. 90
60

and Xj > 0 [0U-MBA-Feb93]

12. Max. Z= 10 xj +

s.t.

3x 2 < 80
,xs >0 [0U - MBA - May 94]

Practice Problems in Duality

1. Min. 3 xy + 4x 2
Subject to 2x 1 + 6 x 2 < 1 6 ; 5 x x + 2 * 2 ^ . 2 0 , xx, x2 > 0
Answer; Max. Z= 16 tfj - 20 t^2»
S.t.
2.. Min. Z = x
S.t. x - 3)> + 4z = 5'; x - 2y < 3; 2^ - z. > 4 ; x , z > 0 &:j; is unrestricted
Answer: Max. Zw = 5 u»j + 3 w^ + 4 w^
S.t 1^2-^2—1' ~ ^ w 1 + 2 t^2 + ^ ^3 - 1
w
4IL>I-W$= 1 ; a>3,1^2 - 0 , i is unrestricted
3. Max. Z=Sa + b + 2c-d
S.t ' 2 a - ^ + 3c + rf=l;a + fc-c + rf=3; a ,b ,c>0 , d\s unrestricted

Answer: Min. Zw = w^ + 3^2

S.t. 2wj + w^ ^ 3 ; — Wj + ^2 - 1» ^wi ~ W2 ~ %
wl + w2 = -l ; wl 8c w2 are unrestricted.
174 Operations Research : Theory and Practice

4. Min. Z = 1 0 x 1

S.t xj + 5 x 2 + x 3 > 1 ;3xx +'^2 ~X2> - 2 ; xj , x 2 , x 3

Answer: Max Z^ = t ^ + 2<x>9

S.t. u^ + 3t^9 < 10 ; bw^ + p2 ^ 6; a;] — a^ - 2; t^j ^

5. Max. Z = x-)? + 3z

Subject to x+y + z < 10 ; 2x - z < 2

2 x - * y + 3z ' ^ 6 ; x,^, z > 0

Answer: Min. Z w = lOrt^ + 2 if 2 + 6 w§

S.t. »! + 2 ^ 2 + 2^3 > 1 ; wl-2v)^ >-l;

^ 2 - ^ 2 + ^ w% - 3 ; 1^1, 1^2 > W 3 - ^

6. M a x . Z = 3 Xj + x 2 + 4 x 3 + 9 x 5 ,

S.t. 4 x j - 5 x 2 - 9 x 3 + x 4 - 2x 5 < 6 ;

2 xx + 3 x 2 + 4 x 3 - 5 x 4 + x5 < 9 ;

Xj + x,2 — 5 x 3 — 7 X4 + 11 xg < 10;

x1,x2,x3?x4,x5 >0

Answer : M i n . Zw = 6 w^ + 9 w^ + 10 tf3

S.t. 4 zfj + 2 1^2 + W3 — ^» - 5 w^ + 5 t^2 + ^ 3 - 0 ;

- 9 wj_ + 4 1^2 ~ ^ 1^3 > 4 ; W\-hw<£-l w<£ >\ \

- 2 tfj + wq, + 11 w3 >9 ; tfj , 1^2 , w 3 > 0

7. Max. Z =

S.t. 4a•+ 4b + 4c + 8d = 21; 3<z + 17* + 80c + 2d < 48 ;

a, b > 0 & c, ^ are unrestricted

; Min. Z w = 21 wx + 48 t^2

S.t. 4 u>! + 3 w% > 6 ; 4 zf j + 17 w^ > 4 ;

4 Wy + 80 w 2 = 6 ; 8^! + 2 ^ = 1 ;

w.2 ^ 0 u>\ is unrestricted.

Simplex Method 175

8. Min. Z ^ = X 3 + X 4 + A;5

S.t. xi - x 3 +X4 + X5 = - 2 ;
^>0(;=l,2, ...5)
Answer: Max. Z^ = + 2 -tf 1 + u»2
S.t. t^j-^2-1' Zfj
- w j + ^2 ^ 1 ; w j ' , % are unrestricted
9. Max. Z = 3x + 2y .

S.t. 2*+;y<5; x+3><3;

x,3>> 0
Answer: x = 2; ^ = 1 , Max. Z = 8; t^j = 1 , w 2 ~ * > ^ m - ^ =
8

Dual is Min Zw = bw^ + 3 ^ 2

S.t. 2wj + u>2 ^ 3 ; z^i + <^2 - 2; w\ $ W2 - ^

10. (i) Use simplex method to maximize

Z = 5x - 23? + 3z ; subject to the restrictions

2x + 2y-z>2;5x-4y<3,y + 3z<5
where x, y, z are non negative variables

(ii) Verify your solution using duality

Answer : x = — ;y = 5,z = 0, M a x . Z~ —
o o
11.
S.t. x+y + z < 7 ; x+ 2^ + 2z < 13 ;
Sx^y +z < 5 ; ' 3 c , ^ , z >0- .

.* Min.

S.t. wy + W2 + 3ix>3 > 2 ; w1

12. Write the dual of the primal problem given below :

Minimize : Z = x± + x2
S.t. 2xj + 4x2 > 4
2xr+^2^3
-xj + 3*2 = 5
x 1 ,^ 2 - 0
176 Operations Research : Theory and Practice

Answer : Max Zd = Aw^ + Sw2

S.t. 2iol + 2w2-ws <1

4tUy + W 2 + 3*^3 <1

n/j, w 2 ^ 0; w 3 un restricted [ J N T U Mech/Prod/Chem 2001/S]

13. Apply simplex to Max. Z : 30z + 236 + 29c

S.t. 6a +• 5fc + 3c < 26 ; 4a + 2b + be < 7 ; a,0tc>-0

Also use duality of the problem from the final table
7 t fi
^nsif^r ; a = 0,fc= — ; c = 0; max. Z = —

2 3 1 6
n - 7
Wl = 0, w2 = y ; mm. Zw = y
14. Give dual formulation of

Max Z = bxx+ 10x 2

S.t. 5 x ^ 8 x 2 <48; 5xj + 3x2 < 48

xt + 8x2 < 48 xh x2 > 0 [ J N T U (CSE ) 9 8 ]

^4n5w^r ; Min. Z = 4 8 ^ ! + 48w 2 + 48t^ 3

S.t. 5wi +5iv2-i-w§> 5; Sw^ + 3w2 + 8r^3 > 10

r^ j , ^ 2 , tt»3 > 0

15. Apply principle of dual to Max. Z = 3 ^ + 2x9

S.t. ^j +x2 > 1 ; xj + x2 < 7 ; xj + 2 x2 ^ 10 ; x? < 3; x1? x2 are n o n - negative

; Min. Z = wi + 7to2 + 10 ^ 3 + 3 ^ 4

S.t. tf j + w2 + ^3 < 3; xuy + w2 + ?^3-+ t^4 > 2 and o^, w2, ^ 3 , -^4^0

: Dual wj = 0 ; w2 = 3, w3 = 0 , w4 = 0 Sc Min. Z = 21

16. Max. Z=2xx+x2

S.t. ^j + 2 x 2 < 10 ; xx + x 2 < 6 ; a^ - x 2 < 2 ; a?! - 2x 2 < 1 ; ^j , x 2 > 0

x± = 4,x2 = 3, M a x . Z = 1 0
Simplex Method 177

17. A diet conscious house wife wishes to ensure certain minimum intake of vitamins
A,B 8cC for the family. The minimum daily (quantity) needs of the vitamins
A, B and C for the family are respectively 30, 20 & 16 units. For the supply of
these minimum vitamin requirements, the house - wife relies.on two fresh foods.
The first one provides 7, 5, 2 units of the three vitamins per gram respectively
& the second provides 2, 4, 8 units of the same three vitamins per gram of the
food stuff respectively. The first food stuff costs Rs: 3 per gram and the second
Rs. 2 per gram. The problem is how many grams of each food stuff should the
house wife buy every day to keep her food bill as low as possible.
(i) Formulate the LPP
(ii) Write the 'Dual' problem
(iii) Solve the primal problem graphically
(iv) Solve the dual problem using simplex method
(v) Interpret the dual problem and its solution.
Answer : (i) Min. Zx = 3#j + 2x2 ,
Subject to 7xx + 2x2 > 30 ; 5x} + _4x2 > 20 ;.2xl +'8x2 > 16 andx , x2 > 0
(ii) Max. Zw = 30 wx + 20 w2 + 16 ws
Subject to Iwy + 5w2 + 2w^ < 3; 2v)\ + 4u/2 + &ws ^ 2; w^ , w2 , u)% ^ 0
5 2
(iii) ^ = - , ^ = 0 ^ 3 = - , Zw = 14

18. Min. Z = - 2 x 1 + 3 2 3

S.t. -2xy+x2 >3; -x2 + x2 >3;

— x2 + 3x2 + x% > — 1 ; xx, x2, x$ > 0

Answer: x± = 0, x9 = 3; x 3 = 0 Min. Z = 9
19. Find the dual of Min. Zp = 4xx + 3 x9 + Sxs
S.t. xY + 2 x2 > 2 , 3 JCJ +x2 + x% > 4 , 4 ^ 3 > 1 , xl + xs > 1 , xx, x2,

Answer: M a x Z ^ = 2w± + 4w2 + zf3 + w^

S.t. wx + 3w2 + w^ < 4 ; 2t^j+^2-^*
3; w^, w2, w§ xv^ ^ 0

20. Prove that the dual of the dual is the primal. Write down the dual of the
following 8c solve them.
Max. Z - 4 Xy + 2 %2
S.t. a?! + x2 > 3; xx - x2 > 2 , x} , ^ 2 > 0
Find dual and hence or otherwise solve the above problems. [0U - MBA Dec. 95]
Answer: x± = 4, x2 = 2, Z m a x = 1 8
178 Operations Research : Theory and Practice

21. Find the dual of the problem Max. Z- 2x-~y

Subject to x+y< 10 ; - 2x+y = 2 ; 4x + 3y > 12 ;x,y>0
Solve the primal by simplex method 8c deduce from it, the solution to the dual
problem.

22. Max. Z =
3
Subject to x± - x 2 < — ; ^i - x 2 > 2, >0
Show that both the primal 8c dual have no feasible solution

23. A dairy has two bottling plants located at^4, 8c B. Each plant bottles up three
different kinds of milk i.e., cow, buffallo and goat. The capacities of the two
plants in number of bottles per shift in a day are as follows.

Plant
Milk
A B
Cow 2,000 1,000
Buffallo 2,000 3,000
Goat 1,000 1,000

Market survey shows that the cow, buffallo and goat milk are required at least
14000, 22000 and 1000 bottles per day. The operating costs per shift of running
plants A and B are respectively Rs. 9000 and Rs. 6000 only. How many shifts
should the firm run each plant per day so that the production cost is minimum
meet to the market demand.

Write the dual of this & give an economic interpretation of the dual variables.

24. Solve by simplex method :

15
Min. Z = — xy+2>x<>
Subject to 3 #i - x 2 - x§ > 3 , xj. — x 9 + x§ > 2 , Xy , x 2 , x% > 0
Write the dual of the above problem. What should be the maximum value of
the objective function of the dual.

25. (i) Min..Z = xl+ 1 0 * 2

S.t. Xj - 5 x 9 > 0 , X\ - 5 x 9 > - 5 & • #i , x 2 > 0

(ii) Max. Z = - 5y
S.t. x + y < 1 , - 5 x - 5 y < - L\J ; x 3 y > 0
Explain how (i) 8c (ii) are related?
Answer : One is dual of the other
Simplex Method 179

26. A company makes three products^, B 8c C out of three materials< * JV> P$

the three products use units of the three materials accordingly.
Pj P2 P$ Unit profit (Rs.)
A 1 2 3 3
B 2 1 1 4
C 3 2 1 ' 5
Available (units) 10 12 15
Determine the product mix to maximize total profit. Solve the primal problem
& write the dual & give geometrical interpretation.
27. Use duality, if any to solve the following LPP.
Min.Z = 15x + \0y
S.t. 3x + by > 5 ; 5x + 2y > 3 ; x,y > 0
5 16 235
A ™- 7
Answer; x - — ,y = — , Mm. Z = —
28. Use duality to Max. Z = 6xj + 5 x2 - 3 x 3 - 4 x4
S.t. 2"Scj + 3 x2 + 2 x 3 - 4 x 4 = 24 ; < 10
x± + x 2 + 2x 3 + 3 x 4 < 15 ;
Xj, x 2 , x 3 , x 4 > 0

^4nszfer : Infeasible solution

29. Solve the dual of the following graphically.

M i n . Z = \0a + 6b + 2c

S.t. -a + b.+ c>\\ $a + b-c>t ,ab,c>0

1 5
Answer : a - —, b = —,c = 0 Min. Z - 10, check answer with graph of dual
Max. Z = x1 + 2x2 S.t. - x j + 3x 2 <10;
Xi + In _^ O , Xi — Xq b 4 ,

Xj , x 2 > 0

30. A firm makes three products yl, B 8c C each product requires production time
in each of three departments as shown.
Time Taken (Hrs/Unit)
Products
Dept. I Dept. II Dept. Ill
A 3 2 1
B 4 1 3
C 2 2 3
180 Operations Research : Theory and Practice

Total time available is 60 hours, 40 hours, and 30 hours, in depts. I, II & III
respectively. If product A contributes Rs. 2 per unit & product B and C Rs. 4
and Rs. 250 respectively, determine the optimum product mix
Write the dual of this problem 8c give its economic interpretation.

31. Find the dual of the following if it exists or otherwise solve it.
Min. Z= 6x\ + 5x2 - 2x3
S.t. x 1 + 3x2 + 2x 3 >5 ; 2x1 + 2x2 + x 3 > 2 ;
4xj - 2x2 + 3x3 > - 1 ; and xl9 x2 , x3 > 0

32. Use dual simplex method to solve the LPP.

Min. Z = - 4#j - 6x2 - 18x3
S.t. xy + 3x3 >3
Xj-x2+x3 > 2 and
xl9 x2,.x$ >0 [JNTU Mech/Prod/Chem 2001/S]

33. Find the dual and solve the linear programming problem :
Min. Z=20xl + 14x2
S.t. xx+x2 < 18
5x! + 3x 2 > 10
4xj + 6x 2 = 10
and xj > 0, x2 unrestricted [ou - MBA Apr/May 94]
34. Find the dual and solve the following LPP.
Min. Z-4*j+x2
S.t. 3x r +2x 2 > 6
- 3x j - x2 ^ - 4
4xj + 5x2 = 10
xx unrestricted and x2 > a [Ou - MBA, May 92]
35. Find the dual and solve the following LPP.
Min. . Z = 10x1 + 8 2
S.t. x! + 2x 2 > 5
2xx-x2 > 12
xy + 3x 2 > 4
xx > 0 and x 2 unrestricted [O u - MBA Aug/Sep. 99]
36. Find the dual and solve the following
Min. Z = x i - 3 x 2 + 2x 3
S.t. xx + 3 x 2 - x 3 > 7
- 2 x 2 + 4x 3 > 12
- 4x{ + 3x2 + 8x3 < 10 [0U - MBA Dec 2000]
Simplex Method 181

37. Solve the dual of the following using simplex method

Max. Z= 3x| + 5x2
S.t. xx + 4x2 < 9

*1>*2 - ° [OU - MBA Apr. 98, Sep. 98, Dec. 2000]

38. Find the dual of following LPP.
Min. Z = 15x1 + 10x2 + 5x3
S.t. xj + 2 ^ + ^ 3 ^ 10
3xj + 5^2 + ^3 ^ 12
4xj + 7x2 + 6x3 = 60
xj > 0, %2 and x$ unrestricted.

39. Find the optimal solution to the following LPP by solving its dual.
Min. Z = 6x1+ 7x2
S.t. 3xj + 9x2 > 3 6
6xx + 2x2 > 24
.2*!+-2x 2 * 16
a
«d x j , x2 > 0 [OU - MBA Nov. 94]

40. Write the dual of the following LPP a n d interpret it

Max. Z =
S.t. %X\ + X9 ~ ^
3^j - ^2 = ^
*!, x2 unrestricted [0U - MBA Apr. 98, Sep. 98, Dec. 2000]

41. Find the dual and solve

Max. Z = 3xj + 5^2
S.t. 4xj + 5x2 < 10
-5xl + 2x2 >-6
5xj + 3x2 =15
and x 1 > 0, x2 unrestricted
42. Find the optimal solution to the following LPP by solving its dual.
Max. Z=S0xx+ 120x2
S.t. xx+x2 <9
20xj + 50x2 < 360
xx > 2
x2 >3
and x 7 > 0 for all7 [ou - MBA, May 91]
182 Operations Research : Theory and Practice

43. Formulate dual of the following LPP and solve it using graph
Max. Z = 2000* + 3000?
S.t. 6x + 9y < 100
2x+y <20
x, y > 0 [0U - MBA Sep. 98, Apr. 99, Sep. 2001]

44. Solve the dual of the following LPP using simplex method verify by graph.
Max. Z - 8 T + 4 C
S.t. 4T+2C <30
2T + 4C <24
' T, C > 0 [0U - MBA July 2000]

45. Find dual of

Max. Z = 4xj + 5x 2 + 4x 3 - 2x4.
S.t. 3xi - 2x9 + 3XQ + X 4 = 5

xY,x2 >0
x3 8c x4 unrestricted [0U - MBA Nov. 94]

46. Find dual and solve

Max. Z = 4xj + b%2 + 6x3
S.t. 2xt + 3x2 + 4x3 < 12
4xj + 6x3 + 7x3 < 14
x1? x2 unrestricted, x3 > 0 [0U - MBA May 95]

47. Find solution for the dual of

Max. Z= 2xj + 5x2
S.t. x t + 4x2 <24
3xx +x 2 < 21
x-i +x 2 < 9
[0U - MBA May 95]
48. Find dual and solve
Max. Z = 3xj + 4x2
S.t. xr+2x2<4

4xj + 5x2 •=' 10

xlfx2 unrestricted [0U - MBA Feb. 93]
Simplex Method 183

49. Solve the dual of

Max. Z = 5x± + 3^2

S.t. 3xl+x2 < 12
xx + 2x2 < 14

*1>*2 - 0 [0U - MBA Feb. 93]

50. Find solution by solving dual

Min Z = 2500 x{ + 2000 x 2 + 500 * 3
S.t. 10x1 + 5^ 2 +x 3 > 2 3
6*1 + 10x2 + 2x3 > 32
Xi,x2,xs >0 IOU - MBA Apr. 94]

Review Questions
1. Give the algorithm of simplex method to solve an LPP. [ECE - 96/OT]

2. Discuss different types of solutions you get in LPP. How do you identify the
occurrence of such solutions in simplex and graphical methods [Mech. 95/C]

3. With a flow chart explain the method of solving LPP by simplex. [Mech. 96/P]

4. With a flow chart, explain penalty method of solving LPP

5. Give step by step procedure to solve LPP by BIG -M method. [EEE - 97/OT/S]

6. Through a flow chart discuss two-phase method of solving LPP

7. Give 2-phase algorithm for solving linear programming problems

8. Discuss the merits and demerits of 2-phase and Big-M methods.

9. Do you think two-phase method is superior to Big-M method to solve LPP. If

so why? If not why not?

10. With reference to simplex, Big-M and two phase methods of solving LPP discuss
the following terms.
(a) Decision variable (b) Slack variable
(c) Surplus variable (d) Artificial variable
(e) Basic variable (f) Non-basic variable
(g) Entering variable (h) Leaving variable
184 Operations Research : Theory and Practice

11. With reference to LP problems discuss the following terms.

(a) Solutions (b) Basic solutions
(c) Basic feasible solutions (d) Optimal solutions
[OU - MBA - D'2000, M 92]

12. Discuss the significance of artificial variables in simplex method of solving LPP.
[OU - MBA - S 99]

13. How do you identify the following solutions in simplex, penalty and 2-phase
methods.

(a) Infeasible solutions JJHJ - MBA - D 90, S, 98, M 92, M 95, J 2000]
(b) Unique solutions [M 95]
(c) Multiple optimal solutions [A 94, S 98, N 94,3 60, S 2001, F 93, M 95]
(d) Unbounded solutions [0U - MBA A 98, M 91]

14. What is meant by degeneracy in simplex? When does it occur? How do you
resolve it? [0U - MBA 90 D, M 92, A 94, F 93, S 2001]

15. With an example discuss the degeneracy in simplex.

16. In the case of tie in entering variable in simplex, how do you resolve? Justify
your answer with an example.

17. How do you convert the non-standard simplex into standard one in the
following cases? Give suitable examples to each case.
(a) Objective functions
(b) Constraint set
(c) Conditions of variables.

18. How do you deal with the following cases in simplex method of solving an LPP.
(a) Unrestricted variable (b) Exact constraints
[0U-MBA-95D, M91, S'01]

19. What is cycling in simplex? When do you get such problem? How do you
proceed in such case?

20. How do you set standard LPP in Big M and 2-phase in the following cases.

(a) Maximization case of three variables with one < , one = and > constraints
and all variables are non negative.

(b) Minimisation case of three variables in three constraints as one '<', one
' = ' and one '>' constraints and all variables are non negative.
Simplex Method 185

(c) Max. case of two variables in two constraints one * = ' and one *>' type
constraints with one of the variables unrestricted and the other
non-negative.

(d) Min. case of two variables in two constraints one ' = ' , one '>' type
constraints and with one of the variables unrestricted and the other
non-negative.

21. What is meant by mixed constraints? [OU - MBA - S 99]

Review Questions in Duality

1. What do you understand by the concept of duality in simplex.

2. Discuss the advantages of duality in respect of simplex method of solutions for

LP problems. [OU - MBA - S 98]

3. Dual of a dual is primal. Verify this statement with an example.

[0U-MBA-N94]

4. Discuss the principles/rules to follow when a primal problem is to be converted

into its dual.
5. Distinguish between the terms of Primal and dual of simplex of LPP.

6. Compare primal and dual characteristics with reference the formulation and
solution method. [OU -MBA - M 91, M 92]

7. What do you understand by shadow price? What is its significance in simplex

method of solving LPP. [OU MBA - S 98]

8. Write step by step method to convert a non - standard primal problem to its
dual.
9. Give the applications of dual problem in LPP.

10. Write a short note on the concept of duality in LPP.

[OU - MBA - 94 D, 95 D, M 91, M 92, S 2001]

11. How do you get the solution of a primal problem from final solution tableau
of dual?

12. Discuss the significance of primal-dual of LPP with reference to game theory.
13. If a primal has m constraints and n variables, how many constraints and variable
will its dual have ? [OU - MBA - A 98]
186 Operations Research : Theory and Practice

Objective Type Questions

1. In simplex table, the value of Z - C: for basic variables will be
(a) negative (b) positive
(c) zero (d) unity

2- The solution of LPP is not an infeasible if

(a) all minimum ratios are either infinity or negative
(b) artificial variables are not replaced
(c) a replacement pf variables is cycle
(d) phase -1 does not yield any solution in 2-phase method

3. Which of the following is a degeneracy in the case of simplex

(a) total supply is equal to total demand
(b) the number of allocated cells exceed the rows+columns - 1
(c) the number of occupied cells is less than rows+columns -1
(d) a tie exists for leaving basic variable.

4. If a primal has unrestricted variable, its dual will have

(a) corresponding variable unrestricted
(b) corresponding constant is requirement type
(c) corresponding constraint is exact type
(d) the constraint is written by splitting the corresponding variable into two parts.

5. Which of the following will have multiple optimal solutions

(a) Z - C: is zero for non basic (decision) variables
(b) cycling of variables in the basis
(c) all slack variables are not replaced and C - Z is still negative
(d) all the artificial variables are not replaced.

6. For a set of 'm' equations in 'n' variables (n>m), the number of solutions will
be
(a)m^ ^ncm
(c) nc (d) infinite
n - m

7. If primal problem has unrestricted variables then its dual will have
(a) > type constraints (b) < type constraints
(c)= type constraints (d) unrestricted variables
Simplex Method 187

8. If the solution for a primal problem is infeasible, its dual will have
solution

(a) multiple optimal (b) unbounded

(c) infeasible (d) unique

9. For 'm dissimilar equations with V variables (n>m), the number of basic
feasible solutions will be
(a)rac (b)nc
n tn

(c) nr (d) infinite

n- m

10. In a simplex table Key Column is determined by

(a) most negative value of C- - Z
(b) most negative value of Z - C;
(c) minimum value of (solution value)/(coefficient of variable in constraint)
(d) none of the above
11. If there is a tie for entering variable, we give preference to
(a) slack variable (b) surplus variable
(c) decision variable (d) artificial variable

12. If a constraint is less than or equal to type, we add

(a) slack variable only
(b) surplus variable only
(c) surplus variable and artificial variable
(d) slack variable and artificial variable
13. If there is a tie for Key column selection, then preference is given to
„ _ variable
(a) slack (b) surplus
(c) artificial (d) arbitrary
14. If minimum ratio is same for a two rows, preference is given to the row having
variable
(a) slack (b) surplus
(c) artificial (d) arbitrary
15. Zj- Cj value for basic variable in a simplex tableau will be
(a) any positive value (b) zero
(c) negative value (d) >0
188 Operations Research : Theory and Practice

16. Simplex tableau yields an optimal solution if

(a) basic variable has C - Z >0
(b) all non basic variables have Z - C <0
(c) all non basic variables have Z- - C->0
(d) all variables have Z- - C<0

17. To improve the simplex tableau, we try to make

(a) basic variable equal to unity (b) key element equal to unity
(c) solution value=0 (d) non basic variable =0

18. In simplex tableau we neglect value for minimum ratio

(a) unity (b) zero
(c) infinity (d) positive

19. In an LPP: Max Z=5xY + 6x2, subject to 2xL + 3x2 > 50 , 4xY + 3x2 > 100, the
objective function of first phase is
(a) 5x| + 6x2 (b) + 5xl + 6x2 - MAX - MA2
(c) +A±+A2 (d) 0.X| + 0JC2 -A\ ~A2

20. The minimum ratio is neglected if

(a) zero (b) negative
(c) infinity (d) infinity or negative

21. In simplex tableau, the basic variables will constitute matrix in

the body
(a) null (b) unit
(c) inverse (d) singular

22. If the coefficient of variables shown in the form of matrix in primal and dual,
then one is matrix to the other
(a) unit (b) inverse
(c) transpose (d) null
23. Coefficients of variables in objective function of a primal problem are
of dual problem
(a) coefficient of variables in objective function
(b) solution values of constraints
(c) conditions
(d) zero
Simplex Method 189

24. If primal has three, constraints, the dual will have

(a) three constraints (b) three variables
(c) three conditions (d) any of the above
25. If a dual has three variables in two constraints, its primal will have
(a) three variables and two constraints
(b) three conditions and two variables
(c) two variables and three constraints
(d) two conditions and three variables

26. The solution values of a dual-solution are of primal solution

(a) solution values (b) zeros
(c) shadow prices (d) negative values

27. If a set of values satisfy all conditions, constraints, it is called

(a) feasible solution (b) optimal solution
(c) multiple solutions (d) infeasible solution

28. For an LPP: Max Z^ = 5 xj + 3 * 2 subject to 2 xx + %2 - 50 > x l + X2 - 3 0

x, > 0, x% is unrestricted, dual is to be written. Which of the following is false
(a) min. Zd - 50 wy + 30 w% (b) 2wx + ii>2<5
(c) wx + w2 > 3 (d) wl>0, w2>Q

29. For a primal Max Z= 50 J^ + 40-x2, S.t. 3 x 1 + 4 x 2 < 6 0 ; 2x 1 + 3 x 2 < 7 0

x y > 0, x2 > 0, the dual is written. Then objective function of dual will be
(a) max. Z = - 50 xx - 40-x2 (b) min. Z = 50 xx + 40 x2
(c) min. Z = - 60 wx - 70 zv2 (d) max. Z = - 60 wx - 70 w2

30. Penalties are used in

(a) Two phase method (b) Big-M method
(c) dual simplex (d) all the above

Fill in the Blanks

1. Dual of a dual is

2. Coefficients of variables in objective function of dual problem are

of primal problem.

3. If primal is to maximise, dual is to

190 Operations Research : Theory and Practice

4. A surplus variable is subtracted and Artificial variable is added for

type of constraint

5. The basic variable will have C - Z• = _

6. Solution values of primal solution are of dual solution

7. If there is no solution in I-phase for a simplex problem in 2 - phase method,

then final solution is

8. If Zj -Cj = 0 for a non basic variable in simplex tableau, we may get

solution

9. The number of basic feasible solutions for system of three equations with four
variables is

10. The set of values that satisfy the conditions and constraint is said be

Answers
Objective Type Questions :
1. (c) 2. (c) 3.(d) 4. (c) 5. (a)
6. (d) 7. (c) 8.(b) 9. (c) 10. (b)
11. (c) 12. (a) .13. (a) 14. (c) 15. (b)
16. (c) 17. (b) 18. (c) 19. (d) 20. (d)
21. (b) 22. (c) • 2 3 . (b) 24. (b) 25. (c)
26. (c) 27. (a) 28. (c) 29. (d) 30. (b)
Fill in the Blanks :
1. primal 2. solution values or constraints of
constraints
3. minimise 4. requirement or >
5. zero 6. shadow prices
7. infeasible 8. multiple or alternate optimal
9. four 10. feasible solution
192 Operations Research : Theory and Practice

CHAPTER AT A GLANCE
Start

Formulate the information

into a TP table

Is it Convert to minimisation type

minimisation type - Multiply all cells by ( -1) or
- Subtract from highest.

Is it
balanced? Create dummy tow/column
(I supply with deficit supply or
I demand and zero transportation cost

Are any Assume M such that M ±a=M

restrictions as transportation cost to
imposed?

Obtain JBFS
by NWOVAM/LCM

solution degenerate? Resolve degeneracy by

allocating * to appropriate cell.

Check optimal ity condition

by stepping stont/MODI method

Compute
associated costs and

Revise the solutions

Transportation Problem 193

5.0 Introduction
Transportation problem is another case of application to linear programming
problems, where some physical distribution (transportation) of resources is to be made
From one place to another to meet certain set of requirements with in die given
availability. The places from where the resources are to be transferred are referred to
as sources or origins- These sources or origins will have ihe availability or capacity or
supply of resources. The other side of this transportation i.e. to where the resources
are transported are called sinks or destinations such as market centres, gojdowns etc.
These will have certain requirements or demand.

5.1 Formulation
Transportation problem is applied to the situations in which a single product
is transported from several origins (say Oj , Oo Om ) to several destinations (say
Dv D2, • • • 0^). Let us assume the cost of transporting a unit product from 0; toD: is
Cf: and the no. of units transported be.x^ Let die capacity of 0; be a; and requirement of
D: be b•-. Then, die transportation problem can be mathematically written as
m n

Minimise (Total transportation cost) ^ = ^ X Qjx*j

Subject to the constraints

M

x
ijij=aij> * = . m (supply or availability constraints)
J
M

. n (demand or requirement constraints)

i= 1
and xy > 0 for all i \
Conveniently, the above can be represented as a tabular as follows.

Destinations
\ To Supply
or
Fnom\.
available
r L^JJJ a
\

cJr*
Origins

r L^.
Demand
or
Requirement K
194 Operations Research : Theory and Practice

Remark :

In the above table, if total supply = Total demand . the

f=[
transportation problem is said to be balanced inmspoliation, otherwise it is
unbalanced transportation.
The above condition is said to be "balance condition* or Mrimcondition*' of T.P.

5,2 Transportation Problem Formulated as an LPP

From the above section we can understand that the TP is a case in LPP. Thus
the given TP can be formulated into an LPP as illustrated hi the following example.
ILLUSTRATION 1

Forumulate the following TPlnto an LPP

Supply
5 6 7 SO
O2 3 7 4 60
S 5 4 40
Demand 30 55 65 150
150
Solution ;
In die above TP,
Let the no. of units transported from Ol toDl=xl[
the no. of units transported from O{ to D2 = xl2 and so on.
Summarily, the no* of units transported from 0; to /) - x-j
where i= l , 2 , 3 a n d ) = 1,2,3
The the total cost can be obtained by summation of the products of the number
of units to be transported from itft origin (row) to jf^ destination (column) with unit
cost given in the corresponding cell (Cy). This cost is to be minimised.
.\ Objective function is
+
Minimize Z=bxu + frr^ + ^x\$ + 3 .^j 7*22
Subject to the constraints
(i) Supply constraints

X
2[ $®
Transportation Problem 195

(ii) Demand constraints

x
\2
= 65 > 0 where i= 1T 2 3 and/ = 1, 2, 3
Remark:

For the above TP, we have rxc = 3x?> — 9 variables and ?' + r = 3 + 3 = 6
equations. Thus it is difficult or time taking process to solve by simplex technique.
Further, it becomes too difficult if there are more number of row or/and columns. The
other easier techniques are available in TP which will be discussed at length in this
chapter.

5.3 Transportation Algorithm: Solution Method

The solution algorithm of transportation problem (T.P) is summarised as
follows.
Step 1: Formulate the Problem in the Form of Matrix : The formulation of the problem is
similar to thai of linear programming. Here the standard form of objective
function is to minimse the total transportation cost and the constraints are the
supply (available) and demand (requirement) to each origin and destination
respectively.
Step 2 : Standardize the TP: A T.P is said to be standard if it is to minimise the total costs.
If the given problem is not standard, it is to be first standardized.

Maximisation Case in Transportation Problem

The T.P. (Maximisation type) is standardised (i.e., converted to minimisation
type) by one of the following two methods.
(i) Identify highest element among all the elements in the cells of transportation
maLrix. Subtract the element of every cell from this highest element and put in
their respective cells, This is now called equivalent cost matrix,
(ii) Multiply by (- 1) to the elements in all the cells of transportation matrix.
Ex: Maximisation :

0\9 &>
2 &$
5 Supply

3 0 -3
0? 63 07 - 93 30

Demand 15 45 40
196 Operations Research : Theory and Practice

Ex: Minimisation :
by method (i) Supply by method fii) Supply
0 7 4 50 -9 _2 -5 20
1 9 12 60 -8 0 +3 30
3 2 0 40 -6 -7 -9 50
Demand 15 45 40 Demand 15 45 40

Step 3 : Check Rim Condition and Obtain Balanced. Transportation Problem,:

5.3,2 Unbalanced Transportation Problem

The rim condition is that the total supply and total demand are equal
i,eM Total supply (availability) = Total demand (requirement)
If the rim condition is not satisfied in the given T*P+, it is said to be unbalanced
1\P. and is therefore to be balanced by creating a dummy row/column with the amount
equal to short fall (or defficiency), of the demand / supply respectively. The respective
transportation are assumed to be zeros.
Example 1: Total supply < Total demand
Unbalanced Balanced
Supply Dt Supply

19 12 L5 20 o, 19 12 15 20

0, 18 10 13 25 IS 10 13 25

16 17 19 50 1G 17 19 50
Dummy
Demand 15 45 40 v?5 (row) 0 0 0 5
Demand 45 40 v100

Example 2 ,-Total Supply > Total demand

Unbalanced Balanced
Dummy
Z>, D2 Supply )3 (column) Supply
19 12 15 20 19 12 15 0 20

09 18 10 13 35 18 10 13 0 25
16 17 19 50 16 17 19 0 50
Demand 15 45 v10540 Demand 15 45 40 .105
100
Step 4: Obtain Initial Basic Feasible bolution (IBFS) :

After getting standard balanced TP, an initial basic feasible solution is obtained
by one of the following methods.
Transportation Problem 197

(i) North West Corner Method (NWCM)

(ii) VogeVs Approximation Method (VAM)

(iii) Least Cost Entry Method (LCEM) or Matrix Minima Method

TJtese methods are explained in the sections to follow in this, chapter.

Step 5: Test the presence of degeneracy and obtain non-dengerate basic feasible solution :

5.3.3 Degeneracy in Transportion Problem

The IBFS obtained (in the step - 4) is said be degenerate if the number of
allocated cells is less than r + c - 1 where r is number of rows and c is number of
columns in the TP matrix.
Degeneracy : No. of allocated cells i.e., n (Cfy) < r + c - 1
Nondegeneracy : No. of allocated cells i.e., n {C^ >r + c- I

If die problem is non-degenerate proceed to step - 6, otherwise resolve the

degeneracy by assuming a minute allocation £ (read as Epsilon) to a suitable cell such
that x-j + e = X" - e = jct-• and e * 0. Then proceed to step - 6.

Step 6 : Optimality test:

Test the non degenerate basic feasible solution for optimality by one of the
following methods.
(i) Stepping stone method

(ii) Modified Distribution (MODI) Method :

If the solution passes the test, then it is the optimal solution, otherwise go to
step - 7

Step 7: Update the solution :

The solution is improved by the given set of rules (discussed in the sections to
follow) till the optimal solution reached.
At every improvement, step 5 and step 6 are repeated

5.4 Methods of Finding Initial Basic Feasible Solution (IBFS)

There are several methods available for finding initial basic feasible solution of
transportation problem. In this chapter we discuss three important methods viz.
North West Corner Method (NWCM), Vogel's Approximation Method (VAM) and
Least Cost Entry Method (LCEM).
198 Operations Research : Theory and Practice

Flow Chart for Solution of Transportation Problem :

Slari

Permit I ate the in form at ion

inio a IP table

Convert to minimisation type

• Multiply all cells by (-]) or
• Subtract from highest

Create dmnmy row/column

with deficit supply or demand
and zcro Ira nsportaI ion cost

Assume M such that M ±a = M

as transportation cost to
the restricted cell

Obtain IBFSby
NWC/VAM/LCM

R eso I v c d egc n enicy b y

allocating e lo appropriale cell

No
Cht:ck optima]ily condition by
stepping stone/VlODI method

Compute associated costs

and print solution

No

Revise iho solutions

FIGURE 5 . 1 :
Transportation Problem 199

5.41 North West Corner Method

This is the simplest method of all methods for finding IBFS of a TP. This
method is independent of the cost or profit of transportation. The algorithm of
NWCM goes as follows.
Step 1 ; Formulation : Given information is formulated in TP matrix.
Step 2 : Slandaradising (Minimisation form): Check whether the given TP is in standard
form. A TP is said to be standard if the objective function is to minimise the
total cost. However, if the TP is to maximise, it may be convened to
minimisation form either of the two methods given below,
(i) From highest among all transportation profits, subtract all profits and put
back in their respective cells (or)
(ii) Multiply profits in all cells by (- 1).
Note : As fhis method is independent of costs or profits this step moy be ignored aho. But be care fid whiU
optimising the initial solution.
Step 3 : Balancing: Check whether the TP is balanced. The balanced TP is that in which
total availability (supply) is equal to total requirement (demand) of origins and
destinations respectively. If the TP is balanced proceed to next step, otherwise
balance by adding the defict amount of availability/requirement to a dummy
origin/destination respectively with unit costs as zeros,
Step 4 : Allocation at North West Corner :
In the given TP matrix identify the cell at north west corner i.e., top left celh
At the starting of the problem it is the cell connecting origin-1 (Oj) to dies tin at ion-1
(Dj), Now check the availability of 0 l and requirement of D\ and allocate as many
units as possible (i.e., minimum of availability of Oj and requirements of Dx) in this
cell and indicate at one of the corners, usually at the top-right of the cell.
Ex:
D{ Z>2 Supply Supply

*13

X
*21 2%

X
*31 ZZ

Demand

If flj <

Now, if availability of 0Y i+e, a j is less than requirement of D li.e.yb[ then allocate

ay units to the cell ( C ^ ) , otherwise (if by ^a^) allocate b\ units to the cell ( C ^ ) . When
an availability or requirement is exhausted encircle it and delete the corresponding
row/column for the next iteration. Also readjust the availability or requirement of row
or column respectively by subtracting the allocated units,
If a i = by both row 8c column will be exhausted and deleted for next iteration.
This leads to degeneracy in T.P. while optimising.
200 Operations Research : Theory and Practice

Step 5 ; Repeat the above step - 4 till all the availability and requirement are exhausted.
Step 6 : Calculate the total costs by summing tip the products of unit cost in the cell with
number of units of all allocated cells. This is the total transportation cost given
by the initial basic feasible solution.
Note : The TP isto reverted to its original matrix before calculating the maximum profit, if it is is converted
to minimisation case earlier Le, in step - 2
Flow Chart :
The flow chart of North West Corner Method is given below.

Stan

I
l;ormulate the in tor mat ion into TR table form |

Convert into minimi sal ion Ibrm by

• Multiplying all cells wilh (-]) or
• Subtract all from highest dement

Isil Ad d d u m m y ro w/c o I u run

balanced? ( I suppK with delicti supply/demand
X demand and zero cosl ol~ transportation

Identify llic Notih Wo si C or ner

oflhc table

Allocate a& many units

as possible to this cell

Delete the row/column exhausted

and readjust other column/row

Arc all the

supply/demand
exhausted*:

Yes

Calculate associated cost Optimise tliL1 solution

by stepping ski no method
and prim solution
or MODI method

Stop

I FIGURE 5 . 2 :
Transportation Problem 201

The above method is illustrated wiiii an example here below.

ILLUSTRATION 2 ——————— - — -

Consider the following transportation problem

Destination
Source Total
Di I>2 o4
1 2 1 4 30
o2 3 3 2 1 50
O3 4 2 5 9 20
Total 20 40 30 10 100

Determine the initial feasible solution (JNTU Mcch-Prt>d-Mechantromcs-Chem-200i)

Step 1 : Formulation : The given problem is already formulated, hence proceed ro step 2.

Step 2 : T h e given problem is to minimise i.e., Standard. (If it is not mentioned, it is

understood as standard TP).

Step 3 : Balancing: Since the row totals (100) equal to column totals (100), it is already
a balanced TP.

Step 4: Allocation by NWCM

ITERATION TABLEAU I:
Ox D» D-A Supply

0,
1 2 1 4
30

50
3 3 2 1

20
4 2 5 9
Demand 40 30 10 00
10'

In the above TP table, ihe northwest corner cell is (C\ j) or {O\ , D±). In this celt,
we allocate as many units as possible. The availability at this source (Oj) is 30, but the
requirement for the destination (D^) is only 20. Therefore we can transfer only 20
units from Oj to Z)j (20 < 30), Therefore O± will have 10 more units to supply to any
destination other than Dj. And since the Z)j requirement is exhausted we delete this
column (i>!) for next iteration while we adjust the supply of 0] from 30 to 10.
202 Operations Research : Theory and Practice

Verdict of I - Tableau - 1 ;
Cell allocated (CY y) or (0x , D^)

No. of units allocated 20

Deletion for next iteration : Dy column { exhausted for D^)
Readjustment for O} row : 30 to 10 (As 20 units are allocated)
Cell cost of transportation : 20 x 1 = 20
ITERATION TABLEAU 2:
T h e above step is repeated
Sup pi
Allocated cell (C 12 ) is (Oh /) 2 )
| 10
o, 2 1 4 © No. of units allocated is 10
Deletion for next iteration is O±
o,. 50 Readjustment for D2: 40 to 30
3 9 1
Cost of transportation from 0^ to Z)9
20 = 2 x 10 = 20
2 5 9
Demand 30 10 X80

ITERATION TABLEAU 3:
Allocated cell (C 22 ) is (O2 t D2)
D4 Supply No. of units allocated is 30
| 30
20 Deletion for next iteration is D<>
3 2 1 Readjustment for O<> is 50 to 20
20
Cost of transportation is 3 x 30 = 90
9 5 9
Demand 30 : 30 10

ITERATION TABLEAU IV:

Allocated cell (C 23 ) is (O2 , Z)3)
Supply No, of units allocated is 20
| 20
Deletion for next iteration is O2
2 Readjustment for D« is 30 to 10
20 Cost of transportation is 2 x 20 = 40
5 9
Demand ^0 10 ,40
10
Transportation Problem 203
ITERATION TABLEAU 5 :

Supply
|JO_
20
5 9
Demand 10 10 ,20

Allocated cells (C 33 and C34) are (O3, Ds) and (O3 > Z>4)
No. of units allocated are 10 and 10 respectively.
Cost are 5 x 10 and 9 x 1G
i.e., 50 and 90 respectively.
All the units are exhausted
IBFS is obtained
Summary :
Supply

°\
| 20

1 2 1 4
30
| 30 | 20
0, 9 50
3 3 1

o.
LUL
4 2 5 9
50

Demand 20 40 30 10

Cost Calculations of IBFS by NWCM

Allocated Cell From To Unit Cost No. of Units Total Cost

CM
°\ ^i 1 20 20

°\ D2 2 10 20
0, 3 30 90
02 2 20 40

5 10 50

cM o. ^4 9 10 90
Grand Total Cost of Trail sport a (ion 310

Total cost of transportation = Rs. 310/-

204 Operations Research : Theory and Practice

ILLUSTRATION 3

A dealer stocks and sells four types of Bicycles namely Atlas, Bharath,
Champion, Duncan which he may procure from three different suppliers
namely Priyanshuf Qureshi and Raju. His anticipated sales for the bicycles for
the coming seasons are 410, 680, 310 and 550 nos. respectively. He can obtain
900 bicycles from Priyanshu, 600 from Qureshi and 560 from Raju at suitable
prices. The profit per bicycle in rupeesfor each supplier is tabulated below.

^^^Type Atlas Bkarath Champion Duncan

Supptier^^^^ (A) (B) (C) (1>>
Priyanshu
21.50 26.00 19.50 21.00
(P)
Qureshi
20.50 24.00 20.00 21.00
<Q)
Raju
18.00 19.50 19.00 19.50
(R)

Formulate the above information as transportation model and obtain initial

solution by North West Corner Rule. \JNTU - ECE/EEE *• 95/OTJ

Step 1 : Formulation
""•^Type Alias Bharath Champion Duncan
Supplier""\ (A) (B) Availability
(C) (D)
Pi"iviuishu (P) 21.30 26 19.5 21 900
Qureshi (Q) 20,50 24 20 21 600
Raju (R) 18 19,5 19 19.5 560
"\2060
Requirement 410 6S0 310 550 1950"\
Step 2 : Standardisation : The above problem contains profit matrix and therefore it is
to be maximised. This is non-standard form. As standardT.P. is to be minimised
an equivalent cost matrix is to be obtained by any of the two methods (i)
Multiplying with (- 1) in all cells (ii) subtracting each cell value from highest
among the cell values. Here (i) Is used, In fact, there is no need for standardising
here, since NWCR is independent of cost or profit. However, for having a
common practice this is earned out.
"\Type Atlas Bharath Champion Duncan Dummy
Supplier"1-^. Availability
(A) (B) (C) (D) (E) '
Priyanshu (P) - 215 -26 -19.5 -21 0 900
Qureshi (Q) -20.5 -24 -20 -5] 0 600
Raju (R) - IS -19,5 - 19 -19.5 0 560
Requirement 410 680 310 550 no 2060
Transportation Problem 205
Step 2 : Balancing : The above problem is unbalanced since the requirement (1950) is
not equal to availability (2060). As the availability is excess, we create a Dummy
column (Say E) with no profit for any supplier and with a requirement of the
deficit i.e. 110,

Alias Bharath Champion Duncan Dummy

Supplier\^ Avaitabiliiy
(A) (B) (C) (D) (E)
Priyansim (P) 2h5 26 19.5 21 0 900
Qmeshi(Q) 20,5 24 20 21 0 600
Raju (R) Itf 19.5 19 19.5 0 5(50
* \ 2U60
Requiretneiu 410 680 310 550 110
2060\

Step 4 : With usual rules we iterate as follows ro get initial solution.

ITERATION TABLEAU 1 :
A B C D E Avl. Verdict :
[410
p -26 - 19.5 -2.1 0 Allocation lo (P, A )
-21.5 490 No. of units 410
Q -20.5 - 21 -20 -21 0 600 Deletion col. A
Adjustment to ? 900 to 490
R - 19.5 - 19 - 19.5 0 560 Cost i s - 2 1 . 5 x 4 10 = -8815
(Negative cosr means profit here}
Req. 080 310 550 110
1650

ITERATION TABLEAU 2 :
A B C D E Avl. Verdict:
| 410 |490 Allocation lo (P, B)
P
-21.5 -26 - 19.5 -2! 0
(490^ No, of units 490
Deletion ; row P
Q -20,5 -24 -20 -21 0 600
Adjustment to £680 to 190
R - 18 - 19.5 - 19 - 19.5 0 560
Cost : - 2 6 x 4 9 0 = - 12740

Req, 410 310 550 110

190 1160

ITERATION TABLEAU 3 :
A B C D E AvL Verdict :
[410 | 490 Allocation lo (Q B)
P - 19.5 ^2! 0 900
-2.1.5 -26 No, of units: 190
| 190 j
Deletion : col, B
Q -20.5 -24
-20 -21 0
410 Adjustment to g a s 600 to 410
R -is" -19.5 - 19 -19.5 0 560 C o s t : - 2 4 x 190 = - 4 5 6 0

Req. 410
J8S6 310 550 110
190 970
206 Operations Research : Theory and Practice

i ITERATION TABLEAU 4 : Verdict :

A B C D Allocation to {Q, C)
L±i£j | 490 No. of units: 310
- 19,5 -21 0 900 Deletion col : C
-2L5 -26
| 190 | 310 Adjustment to Q as 410 to 100
-20.5 -21 0 000
-24 -20 AW 100 Cost : - 20x310 = -6200

R - 18 - 19,5 -19 - 19.5 0 560

Req- 410 680 550 1)0

660

ITERATION TABLEAU 5 & 6 : Verdict for Tableau - 5 :

Allocation to (Q D)
A B C D E Avl.
|490
No. of units : 100
| 410
p 900 Deletion : row Q
-21.5 -26 - 19.5 -21 0 Adjustment to D as 550 to 450
| 190 | 390 | 100
Cost : - 2 1 x 100 = -2100
Q - 20:5 -24 -20 -21 0 j£-K5^. 100 Verdict for Tableau - 6 ;
[450 | no Allocation to (R, D) and (R, E)
R -18 -19.5 - 19
- 19.5 0 110 No. of units 450 and 1 \ 0
Req. 410 680 310 (no) (all units exhausted)
(450) JffiQ 450 Cost : = - 19.5x450 and
Ox 100 =8775 and 0
ITERATION TABLEAU - 7 :
Final Tableau - 7 for FBFS by NWCM
Equivalent Cost Matrix Profit Matrix
A B C D E Avl. B C D Avl.
| 410 490 | 410 | 490
p 900 900
-2L5 - 19,5 0 21.5 26 19.5 21 0
| 190 |31Q ] 100 | 190 | 310 | 100
Q -20.5 -20 0 600 Q 20.5 600
-24 -21 24 20 21 0
| 450 | no | 100 | 110
R - 18 - 19,5 - 19
- 19.5 560 18 19-5 19 560
0 19.5 0
Req, 410 680 310 550 110 2060 Req. 410 680 310 550 110 2060
Profit Calculation
Allocated Cell From To Unit Cost No. of Units Total Cost
P A 2L5 410 8S15
P B 26 490 12740
B 24 190 4560

a C 20 310 6200

a D 21 100 2100

R D 19.5 450 8775

C-> R E 0 110 00
Total Profit 43190
Transportation Problem 207

5.42 VogeVs Approximation Method (VAM) or Penalty Method

This method is a heuristic method and concentrates to minimise penalty cost
In this method, the allocation is made on the basis of the opportunity or penalty (extra
cost) that would have incurred if the allocation in certain cell with minimum unit
transportation cost is missed. The major advantage of this method is that the initial
solution obtained by this method is very close to the optimal solution or most times,
it is the optimal solution itself.
The Algorithm of VAM:

The algorithm of VAM is as follows :

The first three steps will be common as those of NWCM.

Step 1 : Formulation : The given information is formulated in tabular form with unit
costs, supply and demand.
Step 2 : Standardisation; The T.P is said to be standard if it is in the form of minimisation
of total transportation 'cost'. If the T.P has a profit matrix (i.e., to maximise
total profit), it will be converted to equivalent cost matrix by one of the two
methods given below.
(i) Multiply by (-1) to all cell profits,

(ii) Subtract every cell profit from highest profit among all cells.

Step 3 : Balancing; The T.P is said to be balanced if total supply is equal to total demand,
otherwise, it ts balanced by creating a dummy row/column (whichever is
necessary) with unit costs as 'zeros* to each cell and with the deficit
availability/requirement (or supply/demand).

Step 4 : Calculation of Penalties : Calculate penalty for each row as well as each column
by taking difference between least and next least in the row/column. If in a
row/coJumn, there are two least costs (equal and minimum), then penalty is
zero. This penalty indicates the minimum extra cost or penalty that has to be
paid if the allocation is not made at the cell with least cost.

Step 5 : Allocation: Select a row or column that has highest penalty and allocate as much
as possible to the cell with least cost in the selected row or column. If there is a
tie in selection of highest penalty, selection can be made arbitrary,

Step 6 : Adjustments : Delete the row or column for next iteration whose availability or
demand is exhausted/satisfied. Adjust the supply/demand of row/column after
subtracting the allocated units. If a row and column satisfy their supply and
demand simultaneously both will be removed for the next iteration.
208 Operations Research : Theory and Practice

Step 7 : Repetition ; Repeat the step 4 to 6 , till the entire units of supply for origins and
entire units of demand for destinations are satisfied.
Step 8 : Cost Calculation : Calculate the cost by summing up the products of number of
units with unit cost of transportation for all allocated cells. -If the final table is
equivalent cost matrix convened from profit maLrix in sLep - 3, ii is reverted to
profit matrix before calculating the profits.
Flow Chart for VAM :
Flow chart for Voxel's Approximation Method for finding IBFS,
Siart

T
Formulaic the in forma lion iiHo T.P. tiibtu form

Con veil into minimisation form hy

• Multiplying ( I) to all or
• Subtract every element from highest

E&il Create dummy row/column by

balanced? (K supply adding deficit supply or demand
= I deman and ti rinsporUilion ^osl iis zero

Hnd penaEiies by diflcrcnec of least and next

least to all rows and columns. Identify highest penalty
and »o lo I en si cost cell corresponding lo highest penally

Allot;akj as many units

as possible lo tins eel!

Delete row/eoLumn whose supply demand

is cxhiiusiixl and the olher readjusted

Calculate associated cost Optimise the solution

and print solution by stepping stone method
or MODI method

Stop

FiGLIRf. S.i .
Transportation Problem 209

ILLUSTRATION 4

(Standard and balanced TP) [Considerproblem given illustration-2]

Consider the transportation problem

Destin ation
T&tat

1 2 1 4 3O

O2 s S 2 1 SO
4 2 5 9 2O

Total 2O 4O 3O 1O 1OO

Determine the initial feasible solution.

{OH MSc-84, IAS-88,
JNTU B.Tecta Mech, Prod, Mechatronics, Chem - 20017

Solution ;

Ste£ 1 : Formulation : T h e given problem is already formula led, therefore go to step -

2.

Step 2 : Standardisation ; T P is in standard form Le,, to minimise the total transportation

costs.

Step 3 : Balancing; T h e given T P is already balanced since the total supply (30 + 50 +
20 = 100) is equal total demand (20 + 40 + 30 + 10 = 100). Therefore go to
step - S.

Step 4 : Calculation of penalties for rows and columns.

ITERATION TABLEAU 1 :

a. Supply Penalty

1 2 1 4 30 0(1-1)

3 2 1 50 1(2-1)

4 2 5 9 20 2(4-5)

Demand 20 40 30 10 100

Penalty

Penalty for 0^ row : least is J,and next least is also 1, therefore difference is zero.
Penalty for 0% row : leasr is 1, next least is 2 therefore penalty is 1.
Similarly for O 3 ; 4 - 2 = 2.
Also for Dx : 3 - 1 = 2, for D2 : 2 - 2 = 0 for Z>3 : 2 - 1 = 1 and £>4 : 4 - I = 3.
210 Operations Research : Theory and Practice

Step 5 : Allocation :
ITERATION TABLEAU 2 : Verdict:
Allocation at C9MA i.e., [Oo, D44)
Z>i Z>a D$ D4 Supply Penalty ^
No. of units 10, deletion D4 col.
30 0
°] 1 2 1 4 Adjustment for O<y from 50 to 40
["JO 1 Cost : 1 x 10= 10
3 3 2 1 40
20 2
4 2 5 9
100
Demand
20 40 30
Penalty 3
t
Step - : Adjustments : Highest penalty among {0, 1,2 and 2, 0 ,1, 3} is 3. Therefore we
choose D 4 column and we allocate in the cell with least cost i.e. (O2, D^) cell i.e.,
C94, The maximum number of units we can allocate for this cell is 10 units,
since its demand is only 10 though supply is 50. Thus, the demand for D 4 is
satisfied, therefore wre delete this column for next iteration. AIKI for row O2, as
10 units are sent to D^ we adjust it to 40 as fiirther capacity to supply as shown
in Iteration tableau - 2,
Step - 7 : Repeat steps 4 through 6f till atl the centres are satisfied.
ITERATION T A B L E A U - 3 : Verdict :
Allocation at C^ i.e., (On, D9)
Supply Penalty
No. of units 20, deletion O 3
0, 0
1 9 1 30 Adjustment for O9 from 40 to 20
0, 1 Cost ; 2 x 20 = 40
3 3 2 40 Here, highest penalties (2) are in
1 20 tie. therefore we select arbitrarily as
4 2 5 O3.
Demand
20 20 30 90
Penalty

Verdict :
rTERATIOfli TABLEAU 4 :
D D
Allocation at C^ i.e., ( 0 j , D\)
Z)| '2 * Supply Penalty
| 20 1 / No. of units 20, deletion D±
0 0 Adjustment for O\ from 30 to 10
1 2 1 ID
°\ 3 3 2 40
1 Cost : 1 x 20 = 20

Demand
•O 20 30 70
Penally
Transportation Problem 211
ITERATION TABLEAU - 5 : Verdict :
Supply Penalty Tie for penalty is resolved by

®'
arbitration and selected O}
Ox 1 +
2 Allocation (C13) i.e., (OL , D3)
1 No. of units: 10, Deletion : O^
3 2 40
Demand 50 Adjustment forZ)3 from 30 to 20
20 20 Cost= 1 x 10= 10
Penalty

ITERATION TABLEAU 6 :
D2 J>3 Supply Penalty
| 20 20 1
0,
3 2 40
Demand
20 20 40

Here no calculation is required and directly we can allocate as 20 units each to

{Qfy D%) and (O2 , D$) to satisfy all supply and demand with costs as 3 x 20 and
2 x 20 i.e., 60 and 40
Summary :
Supply
| 20 30
°\ 1 9
[ 20
1
| 20
4
50
O2 3 3 2 I
| 20 20
4 2 5
Demand 20 40 30 10

(Check whether sums of the allocated units are same as their respective
demand/supply)

Step 8 : Cost calculation

Allocation cell From To No. of units Unit cost Total

c., o, »i 20 i 20

*>s to 1 10

D2 20 3 fiO

o2 20 2 10
Os
o, O4 10 1 10
20 9 •10

Total 180
212 Operations Research : Theory and Practice

ILLUSTRATION 5

(Maximisation Case and Balanced)

Obtain IBFS of transporation problem whose profit matrix is given below.
Markets
M2 Mi Stock
Godowns G, 4 4 9 25
G2 3 5 8 20
Sales 18 16 11 45

Solution i
Step 1 : Formulation ; The given TP is already formulated.

Step 2 : Equivalent Cost Matrix : As the given TP is in maximisation form, it is io be

converted to equivalent cost matrix. This can be done in two methods [one
method is discussed in illustration - 3, other method is taken here].
ITERATION TABLEAU 1 :

Equivalent Cost Matrix :

NU Stock
Godowns Q 25
b 5 0

Gi
20
2 6 4 1
Sales 18 16 11 45

Highest among all unit profits is 9, All other unit cost are calculated by
subtracting from 9.

Step 3 : Check Balancing: As total stock (25 + 20 = 45) is equal to total sales (18+16 +
11 = 45), the TP is balanced.
Step 4 to 6 : Calculation of penalties Verdict ;
M, M.> Highest penalty is 5 among 5,
3,1,1,1
G
\ .'. Gj is chosen to allocated at C j
5 5 0

G, No. of units : 11
6 A l
Deletion : M 3
Sales 18 16 (u) 45
Adjustment to G] as 25 to 14
Penalty 1 1 1 Cost = 0 x 1 1 = 0
Profit = 9 x 11 = 99
Transportation Problem 213

Step 7 ; Repetition of steps 4 to 6 Verdict :

Highest penalty is 2
My My Stock Penalty
/. G% is seJected; allocation to Co
14 0 i.e., {G9 , My)
5 5
| 16 No. of units: 16

6 4 Deletion : M^
Adjastment to O9 from 20 ro 4
Sales 18 34
Penalty 1 1 Cost = 4 x 16=64
Profit = 5 x 16 = 80

Stock

5
L 14

6
L\ 4

Sales 18

Summary :
Equivalent Cost Matrix Profit Matrix (Original)
M, Stock Stock

LiL LiL 25 Li! LL 25

5 4 4 9
5 0

G2 20 Li. j 16 20
6 4 I 8

Sales 18 16 11 15 Saks IS 1(5 11 45

Step - 8 : Profit calculations :

Allocation cell From To No. of units Lfilit projfii Total profit

G
] Mi 14 4 !5G

c,3 G, Af3 11 9 99

G, M, 4 12

G2 M, 16 3 80

Grand total profit 247

Note ; 77i/j VAM can also be applied to the maximisation problem directh Le., not converting utto
minimisation by taking penalty as highest and next highest and allocating at high profit cell.
However this method may lead to some confusiony therefore students are advhed to follow the above
method only.
214 Operations Research : Theory and Practice

ILLUSTRATION 6

(Minimisation case with unbalanced TP)

Priyanshu Enterprise has threefactories at locations A, B and C which supplies
three warehouses located at Dr E and E Monthly factory capacities are 10f 80
and 50 units respectively. Monthly warehouse requirements are 75, 20 and 50
units respectively. Unit shipping cost (in Rs.) are given as
Warehouse
Factory D E F
A 5 1 7
B 6 4 6
C 3 2 5

The penalty costs for not satisfying demand warehouses D, E and F are Rs* 5t
Rs. 3 and Rs. 2 per unit respectively. Determine the optimal distribution for
Priyanshu using transportation technique. [ J N T U (CSE) 2000|

Solution :
Note : Only IBFS is illustrated here and it may not bepptim&l solution. You have to optimise after finding
IBFS by using stepping stone or MODI method.

Step 1 : Formulation : The given information is formulated as follows:

ITERATION TABLEAU I:
Warehouse
Factory D E F Capacity
A 5 1 7 10
B 6 4 6 80
C 3 2 5 50

Requirement 75 20 50 ,140

Step 2 : Standardisation : Since the given TP has tost matrix, it is in the standard form
i.e,, to minimise of total transportation costs.
Step 3 : Check balancing :

In the above TP, total capacity (140) is less than total requirement (145),
Therefore, thisTP is unbalanced. To balance this TP we create another dummy factory
say K, with the deficit capacity, i.e., 145 - 140 = 5.
The transportation costs usually we assume zeros, but here these are taken as
5t 3 and 2 instead of zeros since these are penalty costs for not satisfying the demand.
Transportation Problem 215

T h u s the reformulated T P tableau is

ITERATION TABLEAU -II:
Warehouse
Factory D E F Capacity
10
A 5 I 7
SO
B 6 4 6
50
3 2 5

K 5 3 2
Requirement 75 20 50

Step 4 to 6 : Penaity calculation and allocation and adjustments :

ITERATION TABLEAU 111: Verdict :
Warehouse Highest penalty for/I i.e., 4
Factory D E F Cavity Penalty Allocation to Cy2 i-e../J to E
No, of units 10
A
I 5O [ I I 7/ Deletion A (exhausted)
80 Adjustment to £ as 20 to 10
B 6 4 6 Cost : 1 x 10= 10
50
C 3 2 5
K 5
5 3 2
50 145
Requirement
10
Step 7: Repeat the above steps :
ITERATION TABLEAU - IV :
Verdict :
Warehouse
Highest penalty Is for Fi,e., 3.
Factory D E F Capacity Penalty
Allocation to C43 i.e., K to F
B 80 No- of units : 5
6 4 6
Deletion : A" (exhausted)
5.
3 2 5 Adjustment : for F from 50 to 45
Cost : 2 x 5 - 10
K 5 3 2
Requirement 75 10 135
45
Penalty 2 1
216 Operations Research : Theory and Practice

ITERATION TABLEAU - V ; Verdict :

Warehouse Highest penalty is for/i.e-, 3.
Factory D E F Capacity Penalty Allocation to C 43 i.e,, KtoF
80 2 No. of units: 5
B Deletion.: K (exhausted)
I Adjustment: for F from 50 to 45
3 Cost: 2 x 5 = 10
Requirement 10 45 130
Penalty 9 1

ITERATION TABLEAU V I : Verdict :

Penalty calculation is not required
Warehouse
Allocations E to D : 25, B to E: 10
Factory D E F_ Capacity Penalty and B to F : 45
B
6
LiL | 45
6
80 2
4
Requiremein 25 10 45 80
Penalty 3 2 1
T
Summary :
Factory D Capacity
10
A 1 7

B 6
LiL LiL 80
4 6
50 50
C 3 2 5

K 5 3 2

Requirement 145

Step 8 : Cost calculation :

Allocation From To No. of units Unit cost Total cost
A E 10 1 10
c21 B D 25 ' 6 150
C
22 B E 10 4 40
B F 45 6 270
C D 50 3 150

c43 K F 5 2 no
Craitd total trailspon cost 620 + 10
= 630
* Total transportation cost is Rs. 620 and penalty for not satisfying the demand of 5 units
to F is 10. Thus total is Rs. 630/-
Transportation Problem 217

5.4,3 Least Cost Entry Method (Matrix Minima Method)

This method concentrates on the cells having least cost. Thus the allocation
entry is to be made in the cell with least cost among all the cells. The algorithm is as
follows:

Step 1: Formulation.: Formulate the information into TP matrix.

Step 2 : Standard TP: If the TP is non-standard i.e., maximisation, convert it to standard

i.e.> minimisation form by either multiplying all profit cells with -1 or by
subtracting from highest profit cell to make equivalent cost matrix.

Step 3 : Balancing : Check whether the TP is balanced by total supply = total demand
and add dummy row/column with deficit and zero costs in the case of
unbalanced.

Step 4 : Allocation : Select the cell with least cost* among all the cost celts and allocate
as many units as possible, and delete the row/column whose supply/demand is
exhausted while the demand/supply of column/row is readjusted by subtracting
the allocated units.

If there is tie in selection least cost, choose arbitrarily or select the one where
maximum allocation can be made.

Step 5 : Repeat: Step - 4 is repeated till all the units of supply/demand are exhausted.

Step 6: Cost Calculation: Calculate the total transport cost by summing up the products
of allocated units with unit cost.

* Least cost entry method is further divided into three types, such as row minima method,
column minima method and overall minima or matrix minima method. In row minima
method we choose least row-wise to allocate while in column minima method we prefer
to take least of each column to allocate. However, these met hoc! are not much popularly
used since these are not close to optimal solution. Therefore overall minima or matrix
minima method only is discussed here.
218 Operations Research : Theory and Practice

Flow chart for LCEM or Matrix Minima Method :

The flow chart for LCEM is given below :
Start

Formulate the information into T.P. table form

Is it Convert into minimisation form by

in minimisation • Multiplying all cells with (-1) or
f cost? • Subtract all from highest element

[sit Add dummy row/column

balanced? (I supply with deficit supply/demand
= I demand and zero cost of transportation

Identify the cell with least

cost of the table

Allocate as many units

as possible to this cell

Delete the row/column exhausted

and readjust other column/row

Are all the

supply/demand
exhausted?

Yes
Optimise the solution
Calculate associated cost
and print solution by stepping stone method
or MODI method

Stop

FIGURE 5.4 :
Transportation Problem 219

ILLUSTRATION 7

(Consider the TP given in illustration - 2 and illustration - 4).

Destination
Source
D, D2 D3 D4 Total
o, 1 2 1 4 30
o2 3 3 2 1 SO
o3 4 2 5 9 20
Total 20 40 30 10 100

Determine the initial feasible solution.

[OU Msc 84, IAS - 88 JNTU Mech, ProdAf echatronks, chem - 2001]

Step 1 to 3 are already done i.e., the given TP is formulated, balanced and
standard.
Step 4 : Allocation : The least cost among all the costs is ' V appearing at three places viz,
at 0] to Z)j, 0\ to Z>3 and O9 to D4 though we can choose arbitrarily the
maximum units can be allocated to (0]t D%) cell Le>, 30< Here, the cell has
supply of 30 and. requirement a] so as 30. Hence both row and column are
deleted for further iteration.
ITERATION TABLEAU I : Verdict :
Chosen at {O^ D$)
Z>i
'1 J
:\
| 30 Allocation to C|^ i.e.. to
I 2 1 4 No. of units = 30
50 Deletion Oj 8c Ds
3 3 9 1
Adjustments : Nil
0, 20
4 2 5 9 Cost : 1 x 30 = 30
20 40 10
Note : We can notice degeneracy while optimisation of the above TP if both row and column get deleted
shtiittlaneously. This degeneracy is explained in the next section under opliw-fsofion ofTP jtith
degeneracy.
Step 5: Repeat step - 4, till all the units of supply/demand are exhausted,
ITERATION TABLEAU II: Verdict:
DA Least cost is * V at {O2 , £>4) i.e., C 2
No. of units = 10
3 3 40
Deletion = D4
0, 20
4 9. Adjustment for 0<> from 50 to 40
20 40 10 Cost= L x 10= 10
220 Operations Research : Theory and Practice

ITERATION TABLEU - 3 : Verdict :

Least cost is '2f
Allocated cell C32 i,e., O3 to D2
40
3 3 No. of units : 20
| 20 Deletion O3
, 4 2
Adjustment to D2 is from 40 to 20
20 Cost = 2 x 20 = 40
20

ITERATION TABLEU - 4 :
It can be directly allocated as :
: 20
| 20 20
C22 or O2 to D2 : 20
3 3
Costs 3 x 20 + 3 x 20 = 60 + 60 = 120

Summary :

O, 30
1 9 1 4
| 20 | 20
50
3 3 2 1
| 20
20
o. 4 2 5 9

20 40 30 10

Step 6 : Cost calculation

Allocated From To Unit cost No. of units Total cost

o, 03 1 30 30

/>l 3 20 60

o2 D2 3 20 60

D4 10 10

D2 2 20 40

G. Total cost 200

Transportation Problem 221

ILLUSTRATION 8

Consider the problem in illustration- 6 which is unbalanced standard T*R The

same is illustrated with this method here.
Solution :
Since first three steps are same, the solution is given from step 4.
ITERATCOIU TABLEAU - 1 Verdict :
D E F Least cost = * 1'
Allocated cell C 12 Le.,A to E
A 10
No. of units = 10
5 1 1
Deletion A
B 80 Adjustment to £ is 20 to 10
6 4 6 Co+st= l x 10= 10
C 50
3 9 5
K 5
5 3 2
75 50 145
10
ITERATION TABLEAU 2 : Verdict :
Least cost is 2 at (C, E) and (K, F)
D E F
Selection made arbitrarily at (C £)
B 80 i.e., C32
6 4 6
No. of units ; 10
C
3
LiL 5 40 Deletion ; E
2
5 Adjustment to C as 50 to 40
K
5 3 9 Cost = 2x 10= 20
75 50 135

ITERATION TABLEAU 3 :
Verdict:
D F
Least cost: 2
B 80 Allocation cell (C43) or (K, F)
6 6 No. ofuntts ; 5
C 40 Deletion : K
3 5 Adjustment to F : 50 to 45
K C o s t = 2 x 5 = 10
5 2
75 125
45
222 Operations Research : Theory and Practice

ITERATION TABLEAU - 4 & 5 : Verdict :

D F Least cost: 3 at (C, D)
:. 40 units are allocated and hence C
B LiL 80 is deleted, then other units allocated
6 6 directly at (B, D) and {£, f) as 35
C 40
(40) and 45 respectively.
3 5 Costs : 3 x 40 + 6 x 35 + 6 x 45
120 = 120 + 210 + 2 7 0 - 6 0 0
(35)

Verdict:

Summary :

A 10
5 1 7
B j 35 | Ah
80
6 4 6
C [ -40
1 10 50
3 2 5
K 1 5 5
5 3 2
75 20 50 145

Cost calculation :
Allocated From To Unit cost No. of units Total cost
c
l2 A E 1 10 10

C21 B D 6 35 210

B F 6 45 270

c31 C D 3 40 120

C E 2 10 20

K F 2 5 10*

630 + 10*
Total transportation cost + penalty cost
= 640

* Penalty for not satisfying the demand (5) of F is Rs, 2 x 5 = Rs. 10

Transportation Problem 223

Now, let us take up another problem with non-standard i.e., maximisation model.
Consider the problem under in illustration 5, to illustrate in this method
ILLUSTRATION 9

Obtain IB FSfor theprofit matrix given below.

Mf M2 M3 Stock
G, 4 4 9 25
3 5 8 20
Demand 18 16 11 45

Solution :
Step 1: Formulation : It is already formulated.

Step 2 ; Standard TP: The given matrix is profit matrix, therefore it is to be maximised.
As standard TP is to be minimised, an equivalent cost matrix is to obtained by
one of the following two methods.

(i) Multiplying by (-1) to all cells.

(ii) Subtracting all profits from highest value among all profit cells to make
equivalent costs. Both methods are simultaneously discussed here below.
Equivalent Cost Matrix :
ITERATION TABLEAU Ha}; ITERATION TABLEAU - 1 (b):
Mi Mn Mo

OR
5 5 0 -4 -4 -9
20 20
6 4 1 -3 -5 -8
t8 1G 11 18 16 11

Step 3 : Balance: It is already balanced since total stock (25+20) = total sales (18+16+11)
Step 4 : Allocation : The least cost among the cells is *zeroJ in 1 (a) [or -9 in 1 (b)] which
appears in C13 cell i.e., from G\ toM^ A maximum of 11 units can be allocated
here. Subsequently, M$ gets exhausted and deleted for further iteration while
stock G[ is adjusted to 14.
224 Operations Research : Theory and Practice

ITEftATION TABLEAU 2 (a): ITERATION TABLEAU 2 {bj:

Mx M2
G
i til OR G!
5 5 0 14 -4 -4 -ft 14
G2 20 20
6 4 l -3 -5 " -a
18 16 (jj) 18 16

Step 5 : : Step 4 till all the uriits are exhausted.

ITERATION TABLEAU • 3(a|: i ITERATION TABLEAU 3(b):

14 OR 14
5 5 ^4 -4
Ztf
6 4 4 -3 -5
18 34 18

Leat cost is 4 tn iteration tableau - 3(a) or -5 in iteration 3{b), which appears in

C22 cell i.e., G2 to M2. So, 16 units can be allocated here since M2 requires 16 only
though G2 can supply 20 units.

UJ - 4(a]: ITERATION TABLEAU - 4(b):

LJL 14 OR LiL 14
5 -4
G2 4 G2 LJL 4
6 -3
IS 18 18 18

S:G M 3 : 11 units, Gj - JV/ 1 : 14 units; G 2 --A^ : 4; G 2 - A f 2 : 16

Summary :
The allocated units are now entered in the respective cells of original TP (profit
matrix). _
1 I u I ~ 1 1 i i
LJL 4 25
4 9

LL 8 20
5

18 16
Transportation Problem 225

Profit Calculation :

Allocated cell From To No. of units Unit profit Total profit

M
\ 14 4 56

C
13 Gl Af3 11 9 99

C2J G2 4 3 12

C99 G, 16 5 80

Total units 45 Total profit 247

AHter ; The maximisation case may also be solved by allocating as many units as possible in the
cell with highest profit. This is a direct method called "highest profit entry method".
Without converting the matrix into equivalent cost matrix. Howeverf this method may
lead to certain confusion and hence not discussed at length.
Note : Answer of IBFS xoith the three methods (NWCM> VAM and LCEM) need not be same.

5,5 Optimisation of Transportation Problem

After obtaining IBFS by one of the three method viz,, NWCM, VAM or LCEM,
we move towards optimality. To find the optimal solution we have two methods

1. Stepping Stone method-

2. Modified Distribution (MODI) method.
These are discussed in the following sections,

15.5.1 Stepping Stone Method

This method is developed by A, Charnes. In this method we use the improvement
index. The algorithm is as follows.
Algorithm :
Obtain IBFS :
Step 1: Obtained Initial solution by any suitable method such as NWCM, VAM, LCEM
etc.
Step 2 : Qieck degeneracy :

Check whether there is any degeneracy. This is checked by the degeneracy

condition that the number of allocated cells is less than sum of number of rows and
number of columns minus one.
226 Operations Research : Theory and Practice

Non-degenerate : n {Cy) > n (r) + n (c) - 1 [or simply n (C-) > r + c - 1 ]

Degenerate ; n (C,y) < n (r) + M (c) - 1 [ or simply n (CtJ) <r + c - 1]

Where n (C^.) is no* of allocated cells in IBFS, n (?) is no. of rows and n (c) is no.
of columns of TP. If there is degeneracy, resolve it, else move to step-3. (Degeneracy
case is explained later).

Step 3 : Calculate Improvement Index : IBFS may or may not be optimum. To check
whether the IBFSJs optimum or not, we calculate Improvement Index, which
indicates possibility of cost reduction by assigning one unit to unallocated cell.

This is done as follows :

(i) Connect one un-allocated cell at a time with all other allocated cells as corners
of a loop. The loop need not always be a square. It may be any shape as given
below,

FIGURE 5.5

Further, the loop should always be closed and should contain either horizontal
or vertical lines only, provided all corners are allocated cells except one whose index
is to be found. Also the loop lines may intersect, but the point of intersection is not a
corner and has no importance. The loop line may go over an unallocated or allocated
cell and such cells have no interference in calculation of index.

Thus after drawing the loop, give positive sign to the 1 in -allocated cell cost and
alternatively negative and positive to the coiners of the cells- Sum up these costs of
cells to find the index.
Step 4 : Revision to Find Improved Matrix ; If all the indices are positive, the optimal
solution is reached (IBFS itself is OBFS). But, if any value is negative, select
unallocated cell with most negative then transfer possible number of units to
the selected cell along the loop and readjust the supply/demand of allocated
cells.

Step 5 : Repeat; Repeat step 3 8c 4 1111 till all the indices are positive.
Transportation Problem 227

Flow Chart for Stepping Stone Method :

Start

Ob la in IBIS by anv method

.such as NWCM/VAM/LCL-M ti

[S tll
so 1 Li I io n dege nc 181J Resolve: decent: mi:y by a I locating
n(r) - n(c) - E such lhal a ± e ~ a.
La ihe appropriate ccll(s)

Improve me ii I Index
Draw lm>p lo eacti unallocated cull with all
other altocattxJ cells
Add and subtract the costs iiheniativclv to ilnd index

Optimal solution is olnainod.

Compute as-socimed cost
and print solution

Yes
Revise salution by adjuslinii and
Iransfemtiii tlie allocated units

FIGURE 5 . 6 :

Consider the problem given in illustration - 2

ILLUSTRATION 10

/># Supply
1 2 1 4 30

o2 3 3 2 1 50
4 2 5 9 20
Demand 20 40 30 10
228 Operations Research : Theory and Practice

Now let optimise this by stepping stone method.

Step 1: Obtain IBFS : (Refer Illustration - 2) the IB IS by North west corner rule is.
i>i Z)9 D^ DA Supply
| 20
1 2 1 4 30
| 30 | 20
0,
3 3 2 1 50

4 2
LlL Li! 20
5 9
Demand
20 40 30 10 100

Step 2 : Degeneracy test by n (Cy) > r + c - 1

No. of allocated cells = 6

[(Olt D$t {Ox, D2), (O2, ^2)' <°2 * &$h (O3, Dz), (O3, Z)4)]

No. of rows = 3
No. of columns = 4
As no. of allocated cells i.e., n (C^) is (6) equal to w (r) + n {c) - 1 Le,,
3 + 4 ~ 1 = 6f there is no degeneracy in the TP.

Step 3 : Optimality Test By Improvement Index ; Loops are drawn for unallocated celts as
given below

Loop for (O[tD$) Cell Loop for (Olt D4)

| 20
£} T
| 20
LiL
1 4 1 1

I io [ ?o | 20
3 1 3 3+* -"<>•?- I i

0,
Li! LiL 1 f° | 10
4 2 5 4 9 -v*-
3-2 = Index ; + 4 - 9 + 5 - 2 + 3 - 2 = -
Transportation Problem 229

Loop for{02j ^1) : Loop for (O^,D{) :

O,
20
_J_IL \ 20
f!
-4^-
1 4 1 4
| 20 ! j 30 | 20
a- 2 1 1
i 3

4 9 5 9
ri ][» LiL
9

Loop for (O2t D4) Cell. Loop for (O$,D2) :

D D2
\

°l 1
20 |io
9 I 4
| 20
LiL
1 2 1 4
1 | 20 [ 30 1 20
3 3 yi 3 T2 1
i LiL !
i-LiL LiL
n« 4 2 5
-* 9 ' 4 5 9
Index :

Of all the above the most negative index is for (O2> Z>4) cell i,e.t - 5. Therefore
we transfer max. possible number of units (i.e., 10) to this cell from {O^ D4).

Step 4 : Revision : When 10 units are transferred from (O3, DA) to (Q2* ^4)- w e have to
reduce 10 units from (O2, ^3) and add at (0%t D%). This addition or subtraction
can be easily understood by sign "on the corner of the loop in the cell. Thus the
revised matrix is,

| 20 LiL 30
1 2 1 4

3
|~~30
3
LiL | 10
1
50
2
4 2
LiL 9
20
5
20 40 30 10

Step 5 : Repeating Steps 2 to 4 : No degeneracy is found and the .indices for unallocated
cells are as follows.
230 Operations Research : Theory and Practice

Unallocated Loop Index

(0,, -2 + 3-2-0

f D4) 2, Z)4)

(0 JP Da)

2-3 + 2-5 = -

. 2 - 3 + 2 - 5 =(-

3. Z)4) -9- 1 + 2 - 5 -

As {O3, D2) shows most negative index we have to transfer 20 units, (this max
possible number of units to be transferred can be found by least allocated units among
the negative cornered cells of the loop). The revised solution is
| 20
30
1 9 1 4

[iL | 30
50
3 3 9 1
| 2Q
20
4 2 5 9
20 40 30 10
Again, the indices in the revised matrix are
for (Ov £3): 0 ; {0}t Z>4) : +4; (O 2 , £>,): + 1 remain same as above tableau,
for (O 3 , D}): Loop is {O3> Drf-* (O5t D2) -^ (O lf D2) -> (Op D}) and index is
+ 4 - 2 + 2 - 1 =+3
for <O3, Ds): Loop is (O3> Z)3) -^ (O 2 , Z>3) -> (O 2 , Z)2) -> (O 3 , Z?2) and index is
+5-2+3-2=+4
for (O 3 P Z) 4 ): Loop is ( 0 3 , Z?4) -^ (O 2 , Z>4) -> (O 2 , D2) -> (O 3 , Z?2) index is
+ 9- 1 + 3 - 2 - + 9
Thus all the indices are positive, therefore optimal solution is obtained.
Solution is
Allocated From To No. of units Unit cost Total cost
CM or(O lT D]) 0. ^1 20 1 20
C{» or {0vD2) £2 10 2 20
C,no\{0^D2) 02 ^2 10 3 30
C a 3 o r ( O 2 , jDg) 0-1 ^3 30 2 60
C e 4 o r (O 2 > Z) 4 ) 0, ^4 10 1 10
C^2 or {Ov D9) O3 20 2 40
G. Total cost 180
Transportation Problem 231

Remark : There is an alternate optimal solution the above problem, which is indicated
by zero index value for unallocated cells. Here (Op Z>3) shows zero. Tims the
solution is

| 20
30
I 9 4
| 20 20
50
3 3 2
| 20
20
4 5 9
5 | 9
20 40 30 10

O p t i m a l ( M m ) cost = 1 x 2 0 + 1 x 10 + 3 x 2 0 + 2 x 2 0 + 1 x 10 + 2 x 2 0
= 180 (Ret Gustation-20)
Consider the TP of maximisation case given in illustration - 5 and 9,
ILLUSTRATION T1

Maximise the TP
M,
4 4 9 25
G2 3 5 8 20
18 16 11 45

Step 1 : The IBFS by VAM or LCEM is

Equivalent Cost Matrix Profit Matrix {Original)
14
25
5 5 0 4 4 9
LL LL Lll 20
6 4 1 8
s
18 16 11 13 16 11

Step 2 : Degeneracy test :

No. of rows (2) +No. of col, (3) - 1 = 4
No. of allocated cells = 4; i.e., n (C-) = n (r) + n {c) -
,\ No degeneracy
232 Operations Research : Theory and Practice
Step 3 : Optimatity test:
¥or(G{TM2) For (G2 -
i n~n
Lil 25
5 5 0 5 0
25
LL I_l6j 20
6 4 1 4
20
18 16 11 18 16 LI
Index .' Index ; + 1 - 0 + 5 - 6 = 0

Since indices for all unallocated cells is positive (or zero), the optimal solution
is attained. And the solution is

Max profit = 4 x l 4 + 9 x 11 + 3 x 4 + 5 x 16 = 24 7 (For detailed calculation refer

illustration 3 or 9)
Remark: The above problem yields multiple optimal solution. This can be
identified by zero index value for non-allocated cells.
The solution is
Mi Mv Mi

4 4 0

3
LlL 20
5 | S8
18 16 11

Profit = 4 x 1 8 + 9 x 7 + 5 x 1 6 + 8 x 4

= 72 + 6 3 + 8 0 + 3 2 = 2 4 7
Note : Fo}' unbalanced case aho, the stepping done method is applied in the similar ?twy after balancing
and obtaining IBFS.

5.5.2 MODI Method

Modified Distribution (MODI) method rs based on the properties of PRIMAL

and DUAL and is simplified or modified version of stepping stone method- The
algorithm is given below.
Algorithm :
Step J : Initial Solution : Develop IBFS by any method such as NWCMf VAM, LCEM etc.
Transportation Problem 233

Step 2 ; Degeneracy Test : Check the presence of degeneracy by n (C^) < n {r) + ?i {c) - 1
where n (C,y) is number of allocated cells, n (f) is number of rows and n {c) is
number of columns. If n {C-) > n (r) + n(c)- } , then there is no degeneracy,
otherwise resolve degeneracy (case of degeneracy is explained in the section to follow).

Step 3 : Shadow Prices Calculation: Find shadow prices of each row and each column with
the formula

Ctj = Ut + Vj

where C,v = Cost in allocated cell

u{= Shadow price of i'h row
Vj= Shadow price of/ column.

As the number of equations will be equal to number of allocated cells and

number of variables are sum of no. of rows and no. of columns* value of one of the
variables is to assumed as zero while calculating the values of u- and v-.

Step 4 : Optiviality Test: Find the cost difference (C^ - ZJJ) for unallocated cells by using
the formula.

where Ct* is cost in unallocated cell and u{ & vj are the corresponding values
(shadow prices) found in step - 3.

Now, if all the values of Cyy - (w( + vp ^ 0 i.e., positive, then the solution is
optimal, othei-wlse improve the solution by step 5.

Step 5 : Identify most negative value of C^ -(&,- + vp calculated in step - 4 among all
unoccupied cells and draw a closed loop with (only vertical/ horizontal lines)
widi this cell as one of its corners while all other corners as occupied (allocated)
cells. [Refer stepping stone method for further details about drawing
loop].Starting at the selected cell as +ve, give alternate +ve and - ve signs to all
comers. Among all the negative sign coiners choose the least allocation and
transfer this amount along the loop. [Add at positive corner and subtract at
negative corner]-

Step 6: Repeat steps 2 to 5 till the C-' - (u7 + vp > 0 for all the unoccupied cells,

Step 7: Calculate the cost [ Profit in the case of maximisation TP ]

234 Operations Research : Theory and Practice

Flow Chart for MODI Method :

Siari

Obtain IB IS by any method

such as NWCM/VAM/LCUM etc.

Resolvu degeneracy by
e such tlial a ± e - a.
to l lie appropriiUc eel its)
No

Calculate u, and v; fshadow prices)

s u c h Ihiii e n s t o f a l l o c a t e d cell {c-j) = u t +• \rj

Calculate cSj'— <u, + \j) tor unoccupied cells

Optimal solution is obtained.

Compute associated cost Slop
and print solution

Revise sol mioa by adjusting and

transferring the allocated unils

FIGURE 5.7 •

Consider the TP of illustration - 2. Let us optimise this by MODI method now.

ILLUSTRATION 12

Optimise the T.P

Supply
1 2 1 4 30
O2 3 3 2 1 50
O3 4 2 S 9 20
Demand 20 40 30 10 45
Transportation Problem 235

Step 1 : Obtain Initat Solution :

Refer ing to illustration - 2, the IBFS for the above TP by North West Corner
Method is as follows.
ITERATION TABLEAU 1 :

1 Supply

0, 30
1 2 1 4
| 30 | 20 50
°2 3 3 2 I
|io |1O 20
4 2 5 9
Demand 20 40 30 10 100

Step 2 : Degeneracy test:

No. of occupied cells n (C(y) = 6

No. of rows i,e+? n (r) = 3
No. of columns i.e., n (c) = 4

Tliere is no degeneracy.
ITERATION TABLEAU 2 :
D4 Supply Shadow price («,)
| so
LiiJ 30
a, = 0 (assumed)
1 2 1 4
| 30 | 20
50 Uy = 1
3 3 2 1
| 10 | TO
20
4 2 5 9
Demand 20 40 30 10
Shadow Price (v:) v^ = 1 v^ - 2 ^^ = 1 ^4 = 5

(Assumed) For Occupied Cells Assume C,-. = «(- + v.-

= 2 , ?/

= 5, W3 + v^ = 9

W e get a b o v e 6 e q u a t i o n s with seven variables u]r u?, u% a n d u j , x^2, and

T h u s o n e of t h e variables ( s a y u j) is a s s u m e d t o b e zero.
236 Operations Research : Theory and Practice

Then all odier values can be found as mentioned in the above tableau.

Step 4 : Optimally test: T o check C^ - {u{ + vj) > 0

ITERATION TABLEAU 3 ;
D} ZXj Z) 3 D4 Supply
| 20 0 - I
30 H.J = 0
1 9 1 4
+1 | 50 | 20
O2 50 ii.2=l
3 3 2 \
- 1 -4 ho
20 u*- 4
4 2 5 9
Demand 20 40 30 10
% = \ v4=5

Sample Calculations :

for (Ov D 3 ) cell C{j = 1 , u^ 0, tj = 1

for (O 2 , ^ 4 ) cell C 2 4 = l , w2 = 1, ^ = 5

F o r ( 0 3 , Z)j) 4 - ( 4 + . l ) = - 1
and similarly for other unoccupied cells
Step - 5 ; Revising the Sohdion: From the above calculations, we find most negative value
of c±j - (w; + Vj) for (t?2 - ^4) c e ^ as - 5, We start constructing closed loop from
this cell with o t h e r o c c u p i e d cells i.e., (O2, D4) -> (O2> ^3) -> (°3> ^3)
—^ (O 3? D4) and on the corners of this loop we give + , - alternatively from
(O2» ^4) onwards. Thus (O 2 , ^4) and (O 3 , Z)3) get addition (+) while (0%, D$)
a n d (O3 , ^4) get subtraction (-). Among these negative marked cells,
{O3, O 4 ) has least allocation as 10 and thus direction is form this cell. (10 units
are transferred upwards). Revising the allocations in this way, we get.
ITERATION TABLEAU 4 :
Dx D2 D3 Z>4 Supply
| 20
Li! 30
1 2 1 4
| 30 \ 10
50
3 3 2 1
| 20
20
4 2 5 9

Demand 20 40 30 10 100
Transportation Problem 237

Step 6 : Repeal steps 2 to 5 :

ITERATION TABLEAU 5 :
D
\ D4 Supply
| 20
Lit 0 +4 30 0
1 2 1 4
+ 1 50 1
3 1
-1 +5
20 4
4 +
9 5" 9
Demand 20 40 30 10
1 9 1 0

AU-6:

*>i a. Supply u
20 |_20_ 0 +4 30 0
o, 1 2 1 4
+1 | so 50 1
3 3 2 1
+3 +4 +9 20 0
4 2 5 9
Demand 20 40 30 10
v- 1 ? 1 0

In the above solution, the values of Cf - - (w, + v:) > 0 for all unoccupied cells.
Therefore the optimality is reached.
Step 7: Cost calculations :
The solution is tabulated as follows.

Cell Front To Unit Cost No. of Total Cost

Units
C u oi{Ol Z>|) 1 20 20

D[ 2 10 20

Cyg or (O«2» ^2) o, 3 10 30

Cq^ or (Oo> ^ S ' 2 30 60

C94 or (Oy, Z)+) 1 10 10

C^
; *> or {D^, D^) 2 20 40

Total Cost 180

Note -• This problem yields alternaie optimal solution. Refer illustration 20 under section 5.11.1
238 Operations Research : Theory and Practice

I L L U S T R A T I O N 13

Describe the transportation problem. Use Vogel's approximation method to

obtain an initial feasible solution to the following product distribution problem*
Transportation Costs
Distribution Centres

Factory
Factory Mumbal Bangalore Delhi Chennai
Capacity
Kotkata 6 5 8 8 30
Ranch! 5 11 9 7 40
Ahmedabad 8 9 7 13 SO
Centre Demand 35 28 32 25

Solution: [JNTU (EEE/OT) 98]

The problem has been solved using VogePs approximation method for IBFS.

Step 1 : Standardisation: Since this problem is a minimization case, it is in standard

from.

Step 2 : Balance Checking ; T h e T P is balanced since the total capacity is equal to total
demand.

Step 3 ; T h e next step is to determine penalties for each row and each column. A penally
is the difference beween the least cost and the next least cost in corresponding
row (or column), for e.g. T h e penalty for the first row i.e,, corresponding to
Kolkata is 6 - 5= t.

Step 4 ; Choose the maximum penalty among all the penalties which have been
determined in the previous step.
Factory Mumbai Bangalore Delhi Chennai Capacity Penalty
Kolkata | 28
6 '5 8 8 2 1
Rancht
5 11 9 7 40 2
Ahmedabad
8 9 7 13 50 1
Demand
35 32 25
Penalty 1 4 L 1
T
(Here 4 is the maximum penalty)
Transportation Problem 239

Allocate maximum possible units to the cell whose cost is minimum in the row
or column selected (corresponding to max. penalty). If the capacity or demand is
exhausted, delete the row or column for next iteration. Prepare the next iteration
table with revised capacities and demand.

After the first 4 steps, we can derive that the allocated cell is Kolkata to Banglore.
T h e number of units is 28 units. T h e deleted column is Banglore column and revised
capacity of Kolkata is 2.

Repeat the above 4 steps and prepare the next iteration tables.
ITERATION TABLE 3 :
Factory Mumbai Delhi Chennai Capacity Penalty
Kolkata
6 8 8 2 2
Ranchi j 35 2
5 9 7 5
Ahmedabad
S 7 13 50 1
Demand
35 32 25
Penalty
1- 1 1

Inference: Allocation to +1 Ranchi to Mumbai" cell-35 units. Delete - Mumbai column,

Revise capacity of Ranchi Row to 5.
ITERATION TASLE 3 :
Factory Delhi Chennai Capacity Penalty
Kolkata
8 8 9 0
Ranchi
'9 7 5 2
All me dab ad p6
7 "" ' 13 18 6
Demand • ('32 i
25
Penalty
1 . J

Inference: Allocation to "Ahmedabad to Delhi" cell -32 units, Deletion Delhi column and
capacity Revised to Ahmedabad Row as 18.
240 Operations Research : Theory and Practice

ITERATION TABLE 4 :

In this case since there is only one column, penalties need not be calculated.

Chen n ai Capacity
Kolkata
8 2
Ranchi
7 5
Ahmedabad 18
13 18
Demand 25

The obvious allocations are

Kolkata LO Chennai cell -^ 2

Ranchi to Chennai cell -» 5
and Ahmedabad to Chennai cell -» 18

Now the IBFS is

Mumbai Bangalore Delhi Chennai Capacity
Kolkata
6 5
| 28
8
\JJ 30
8
Ranchi
5 U 9 7 •10

Ahmcdabad | 32
8 9 7 13 50
Demand
35 28 32 25
Moving Towards Optimality :

First we have to check the non -degeneracy (tt>r + c-l)

No. of allocated cells (say n ) = 6

No. of rows {r) + No. of columns (c) - 1

=4+3-1=6

Hence there is no degeneracy as n = r + c - 1 = 6 .

Now calculating u} and v- values using allocated cell cost (c-) - u{ + v .

Wliere u- and v^ are shadow cost variables of ith row andj column respectively.
And then calculate c^ - ( u, + Uj) where C^ is cost in nonallocated celL
Transportation Problem 241
ITERATION TABLE 5:
Mumbai Bangalore Delhi Chennai Capacity
Kolkata +0 | 23 +6
6 5 8 8 30 0
Ranchi \JL +7 +8

5 p-n v H 40 -1
Ahmedabad 5
• *
^ ,. 50
8 J •-
Demand 5 ;
35 28 32 25
7 ; =7
6 5 2 8 13 ;
: ^ j there are 6 equations, but seven variables, one of them is assumed to be zero (u\ = 0 J

As there is more negativeness existing at Ahmedabad to Mumbai cell, the table

is to be revised by transfering maximum possible units along the loop line.
ITERATION TABLE 6 :
Repeating the step as same as the above in Iteration Tableu 5. (No degeneracy as
- I =6).
Mumbai Bangalore Delhi Chennai Capacity
Kolkata
6 5
| 28 +3
8
Li_ 30
0
8
Ranchi +7 -1
5 11 9 7 40
Ahmedabad +2 32 +3 2
8 9 7 is 50
Demand
35 28 32 25
+0
6 5 5 8
As Ctj > ut + Vj for all non allocated cells, the optimal solution is reached. The
solution is as follows :
From To No. of Units Unit Cost Total Cost
Kolkata Bangalore 28 5 140
Kolkata Chennai 2 8 16
Ranchi Mumbai 17 5 85
Ranchi Chennai 23 7 161
Ahmedabad Mumbai 18 8 144
Ahmedabad Delhi 32 7 224
G, Total Umta 120 G. Total Cost 770
242 Operations Research : Theory and Practice

ILLUSTRATION 14

Solve the following transportation problem, u&> Find the optimal solution,
where the entries are cost coefficients.
Destination
From To Availability
1 2 3 4
1 15 0 20 10 SO
Origins 2 12 8 11 20 50
3 0 16 14 18 100
Requirement 30 40 60 70 200

[3HTU (EEE/OT) 97/S1

T h e T P is solved by using Voge r s approximation method for Initial Basic

Feasible Solution.

Step 1 : Since the T P is a minimization case, it is in standard form.

Step 2 : Given T P is balanced since the total availability is equal to total requirement.
Step 3 : T h e penalties corresponding to each row and each column are determined.
(Penalty is difference between the lowest number and the next lowest number
of a row or column.)
Ex: H i e penalty corresponding ro \st row is calculated as 1 0 - 0 = 1 0
(ITERATION TABLEAU 1:

To Avrr! Penalty
From

10
15 0 20 10 50

12 8 I] 20 50
| 30
0, 14
0 16 14 IS 70

Required
40 60 70 200

Penalty 8 3 8

In thisTP the max. penalty is 14, hence 30 units are allocated to the cell having
minimum cost i.e., 0 of row 0% and consequently 30 units are allocated (in this case)
hence the \st column is deleted and Vd row availability is revised to 100-30^70.
Transportation Problem 243

T h e above steps are repeated for successive iterations.

i ITERATION TABLEAUX;
To Avail Penalty
From

0 20 10 10

8 It 20 50

16 "' 14 18 70
Required '40 r 60 70 170

Penalty 8 3 3
a no
The allocation is from Oj toZ?^ ftd - of units allocated are 40 units. Revised
availability for Oy row is 10.
i ITERATION TABLEAU 3 :
To Avail Penalty
From
10
10

H 20 50 9

14 18 70 4
Required
60 60 130
Penalty

Allocation is in cell to Z>4) and equal to 10 units, delete Oj and revise

requirement of D 4 as 60
i ITERATION TABLEAU 4 :
Avial Penalty
| 50
o2
n 20 0 9 <-

14 18 70 4
Required 0 10 60 120

Penally
3

Allocation is from O9 to D% and is equal to 50 units; Delete O? and revise

requirement of D% as 10
244 Operations Research : Theory and Practice

JTEP4TI0N TABLEAU 5 :

1 10 | 60
14 18

Summary:

15 0 20 10

12 IS n 20
| 30 [ 60
0 16 14 18
T h e total cost incurred is,
10 x 10+ I I x 5 0 + 14 x 10+ 1 8 x 6 0 = 1870

Step 5; Checking for Optimality: T h e given problem is non-degenerate (no degeneracy)

as number of allocated cells n {c^ ) = 6

No. of rows (r) = 3 , No. of columns (c) = 4; r + £--1 = 3 + 4 - 1 = 6

[ .-. n (Cy:) ^ r + c - 1 ]
Now, equations with allocated cells as Cj,r = w?: + v.
We get,

=0

Assuming ttj = 0 t we have ^2 = 5 , ^3 = 8 ,

vj — — 8 , vo - 0 , ^3 = 6 and ^4 = 10
Now checking c^ - {u{ + Vj) > 0 for non allocated cells,
forcn 1 5 - ( - 8 + 0 ) = + 23; forc 13 20-(0+6)=+14

for<?24«)-(5+10)=+5, and forr 32 16-(8+0)=+ 8.

Since alJ cty' - (w: + v-) are positive, the obtained solution is optimal.
Hence the minimum cost = Rs. 1870/-
and allocations are
01 to D2 = 40 units ; Ox to Z)4 = 10 units
02 to D3 = 50 units ; O3 to Z)j = 30 units
0$ to Z>3 = 10 units and O3 to D4 = 60 units
Transportation Problem 245

ILLUSTRATION 15
An oil company has got three refineries P> Q and R and it has to send petrol to four
different depots Af B, C and A The cost^ of shipping 1 gallon of petrol at the
refineries are given in the table. The requirement of the depots and the available
petrol are also given. Find the minimum cost of shipping after obtaining an initial
solution by VA&I.s .
Depositf
Refineries Available
B C D
P 10 12 IS 8 130
Q 11 11 9 10 150
R 20 9 7 18 170
Required 90 100 140 120 450
| JNTU (Mech) 99|

The given problem is

Step 1: Minimisation case of TP -^ cost minimisation (standard form).
Step 2 : Balanced TP model -> Requirements = Availability (450 = 450)
Step 3: We find the penalties and start allocating the row or column which has
maximum penalty.
i ITERATION TABLEAU - 1 ;
Refineries Depot
A B C D Available Penalty
| 120
10 12 8 130/10 2

Q LI 11 9 10 150 1

R 170 2
20 9 7 IS
Required 90 100 140
9 9
Penal ty

Step 4 : Remove the column D as it is filled to its requirement and start allocating to
other columns & rows.
ITERATION T A B L E A U - 2 :
Availability Penally
P 10 2
10 12 15
Q 150 2
11 11 9
R j 140
20 9 1 170/30
Required 90 100
Penalty 1 9
246 Operations Research : Theory and Practice

Step 5: Column C is neglected as it is filled to its requirement.

ITERATION TABLEAU - 3 :
A B Availability Penalty
P 10
10 12

150 0
11 11

20 -' 11

Required 90 100/70
Penalty 1 2

Step 6 : Row R is neglected

ITERATION TABLEAU 4 :
A B Penalty
1 10
10 12 10

Q
I so [ 70
11 11 150

90 70
Penalty 1 1

Step 7 : Finding the Costs of an Allocated Cells :

A B C D
P + 11 +7 | 120
- 10
10 12 15 8
| 80 | 70 0 +1
Q
11 11 9 10
R + 11 | SO | 140 + 11 =9
20 9 •7 18

Initial solution by VAM itself is optimal

Since the values c^ - ( ui + V:) > 0.

= 10 x 10 + 11 x 80 + 11 x 70 + 9 x 3 + 8 x 120 + 7 x 14 = Rs. 3960,

Minimum cost = Rs. 3960.
Note : This problem yields an alternate optimal solution.
For hints, Refer illustration -20 under section 5. 11.1 and try for another solution
Transportation Problem 247

ILLUSTRATION 16
Solve the Transportation Problem whose unit cost matrix^ supply and demand
are given below:

Supply
7 7 10 5 11 45
o2 4 3 8 6 13 90
Os 9 8 6 7 5 96
o4 12 13 10 6 3 75
Os 5 4 5 6 12 105
Demand 120 80 50 75 85
iJNTU(Mech)98/s|

As the matrix is cost matrix, theTP is to be minimised i.e., hi the standard Form,
hence we procede directly.
D, Supply
45
7 7 10 5 11
90
4 3 8 6 13

9 8 6 7 5
75
12 13 10 6 3
105
5 4 5 6 12
Demand 120 80 50 75 ,411

As die D e m a n d (410) ^ Supply (411), the given T P is unbalanced. H e n c e we create

a dummy destination D§ with a d e m a n d of I unit so as to make d e m a n d =supply.
Obtaining IBFS by Vogel's approximation :
1 ITERATION TABLEAU 1:
*>, D* A* A* Supply Penalty

01 7 7 10 5 11 0
LLJ 45/44 5

90 3
4 5 8 6 13 0
96 5
9 8 6 7 5 0
75 3
O4 12 13 10 6 3 0
105 4
o., 5 4 5 6 12 0
Demand 120 80 50 75 85
Penalty 1 1 1 1 0 410
248 Operations Research : Theory and Practice

* (Penalty is difference of least and next least of respective row/column) Highest

penalty is for 0^ therefore one unit is allocated to the cell (0 t , Z)6 ) and all the
units in D 6 are exhausted- Hence Z>e is eliminated for next iteration while
supply of 0v reduces to 44 units.
ITERATION TABLEAU 2 :
Supply Penalty

0, 44 2
7 7 10 5 11
90 1
o, 4 3 8 6 13
96 1
0,
9 8 6 7 5
|:75
1? 13 10 6 3
105
o. 5 4 5 6 12
120 80 50 75
Demand 10 335
Penalty 1 1 1 2

Allocation to (0$ , X>- ) — 75 units

Demand For D§ becomes 10, O± is deleted,

ITERATION TABLEAU 3 :
Supply Penalty

44 2
7 7 10 5 u
90
4 3 8 6 13

9 8 6 7 5 86
105
0*
5 4 5 6 12
80 50 75
Demand 325
Penally 1 1 1 1

Allocation i i , Z)=) - 10 units

Supply of O3 reduces to 86, D5 is removed.

Transportation Problem 249
ITERATION TABLEAU 4:
D» Supply Penalty
| 44
7 7 10 • 5 • • • • . 0
90
4 3 8 6
86
9 8 6 • " • ' ' ? •

0R 105
5 4 5 6
Demand 120 80 50
31 281
Penalty 1 1 1 1
Allocation (Oj , D 4 ) - 44 units, deletion - O}.
Demand reduction for D4 to 3 L

i ITERATfOEU TABELEAU 5 :
Supply Penalty
": j 80.
4 • • :
. - - 3 - : - S 6 10
86 1
9 •'..•;8'- 6 7
105 1
:
5 • A y.- 5 6
130 50 31
Demand
201
Penalty 1 1 1 0

Note; Highest penally is choosen orbitrarily.

Allocation (O2 , D 2 ) — 80 units, Deletion - j
Supply reduction of O2 to 10.

ITERATION TABLEAU 6 :

*i Supply Penalty

®
: 2
,
"•6.-••"!:

S6 1
9 6 7
105 0
5
5 5 6
50 31
Demand
no 191
Penalty I 1 0
Allocation (O2,D\)-— 10 units , Deletions —>
Demand reduction on/)i to 110
250 Operations Research : Theory and Practice

D
\ A* Supply
86
9 6 7
| 105
.^.
5 5 6

Demand ^ ,_. 50 31
86
Penalty 4 1 1

Allocation (Ob t D{ ) Deletion — 0 5

Demand reduction of D} toO 5 .
At
| 30
LiL 86
9 6 7
50 31

Allocation (0 5 , Z)] ) — 5; (O3 , Z)3 ) — 50,

(O^ , JD,| ) — 31 units.
D, Da Dx DA DK

| 44 LL 45
7 7 10 5 11 0
LiL | 80
3 8 6 13
90
4 0

9 8
| 50
6
liL LiL 0
96
7 5

12 13 10 6
LiL 0
75
3
[ 105 105
5 4 5 6 12 0
120 80 50 75 85 1 411

T r a n s p o r t a t i o n cost b y V A M i s ; 5 x 4 4 + 0 x l + 4 x l + 3x80 + 9x5

+ 6 x 5 0 + 7 + 3 1 + 5 x 10
+ 3 x 7 5 + 5 x 1 0 5 - 1862

Optimally Test:
Degeneracy test: No, of allocated cells = 10 ;
No. of rows (r) = 5 ; No. of columns (c) = 6; r + c - 1 = 10
No. of allocated cells (10) > r + c -I {i.e 10); so no degeneracy.
Transportation Problem 251

D, £>, D, D,
0 +1 +6 J 44 +8 45
0,
7 7 10 5 11 O'~
:
80 + 7
IJL +4 + 13 4-3 90 -3
4 3 S 6 13 0

0,
9
0
8
| 50
6
111 5
-2 :
o+
" 7
+5 +7 +6 +1 0 75 0
12 13 10 6 3 0
1Q5 | 0 4-3 +3 + 11 +2 105
5 4 5 6 12 0
120 80 50 75 85
7 6 4 5 3
&9 + V ] = 4
-9; &3 + v5 = 6
U3 + v 4 = 7 ; = 5; =7
D,
0 +1 +6 +3 +2
0, 45
7 7 10 5 11 0
+7 +5
Li! | 80
3 8 6
4-3
3 0
90 -3
4 0
LJL | 30
LJL 96
9 8 6 7 5 0
+5 +7 +6 +1 +2
75
12 13 10 6 3 0
105 0 4-3 +3 +U 4-4
105
5 4 5 6 12 0
120 80 50 75 85
7 6 4 3

From the above tableau C»' - {«,; + vy) > 0 Le*t Non-negative for all
non-allocated cells, the optimal solution is obtained as follows,
Cell Unit cost Allocated units Total cost
(Rs.)

(OlfD4) 5 45

(O 2 ,Z>|) 4 10 40

(O2 , Z>2} 3 80 240

9 5 45

(Os , D3) 6 50 300

252 Operations Research : Theory and Practice

7 30 210

( O , , Dr>) 5 10 50

0 1 0

co.,, u , > 3 75 225

0 105 525

Grand Total Rs I860A

The total minimum transporation cost (optimum) is Rs, 1860/- and one unit in
the supply of origin -3 is not transported to any destination (unutilized resource).
Note : The above TP yields multiple optimal solutions which will be explained in late\r sections. Students
may try.

ILLUSTRATION 17

Find the optimal solution to the Transportation Problem given below;

Warehouses
Wt w2 w4 Supply
Ft 14 25 45 5 6
Factory F2 65 25 35 55 8
F, 35 3 65 15 16
Demand 4 7 6 13 30 (total)

[ JNTU (CSE) 971

Solution i
Ware House
Factory w2 W, Supply
14 25 45 5 6

65 25 35 55 8

35 3 65 15 16

Demand 4 7 6 13 30

Step 1 : As it is minimization case, the TP is in standard form.

Step 2 : The given TP is balanced as total supply = Total demand.
Step 3 : Finding IBFS by Vogels approximation method.
Transportation Problem 253
ITERATION TABLEAU - 1 :

wx wA Supply Penalties

9
14 25 • 45 5 6
10
65 25 35 55 8

35 3 65 15 9
Demand 4 (7) 6 13 30
Penalties 21 22 10 10

i ITERATION TABLEAU 2:

W, Supply Penalties
fi 9
.14 45 5 2

65 35 55 8 20

35 65 15 9 20
Demand 6
30
Penalties 21 10 10
T
ITERATION TABLEAU 3 :

Supply Penalties
40
45 5
20
35 55
50 f-
65 15
Demand 6

Penalties 10 10
i ITERATION TABLEAU 4 :

wA Supply Penalties
•U_ 40
45 5
UL LJL 20
35 55
Demand 6
2
Penalties 10 50
254 Operations Research : Theory and Practice

IBFS TABLEAU:
WA Supply
4 6
14 25 45 5

63 25 35
LL 8
55
16
35 3 15
Demand 4 7 6 13
Initial basic feasible solution = 14 x 4 + 3 x 7 . + 6 x 3 5 + 2 x 5 + 2 x 5 5 + 9 x 1 5 = 5 4 2
n {Cjj) = n (r) + n (c) - 1 i.e., 6 - 4 + 3 - 1, H e n c e n o d e g e n e r a c y

Optimisation :
w
VK> \

14
1 4 + 32
95
+GQ
45 5
2 H, = 0

+1
65 25 " 35
1^ , = 50

+ 70 + 70 .= 10
35 3* 65 * 15
=- 7 =+5

j = 35 ; te^ + U4 = 55

9 = 3; w-3 + v^ = 15
"+32 +42

+ 19
14LjL 25 45 6 wt = 0
2;
+ il
65 25 55 8

35
L5- 65
LLL 16
3 iu=10
15
4 7 6 13
: = 14 v2 = -7 ^ =3

it, y + v y = 1 4 ; 2/ j + 1/4 = 5

&2 +• 7^2 = 25 ; ut + tJ3 = $£

Optimal solution
= 4 x 1 4 + 2 x 2 5 + 3 x 5 + 6 x 3 5 + 2 x 5 + 11 x 15 = 506
Transportation Problem 255

5*6 Degeneracy in Transportation Problem

A transportation problem is said to be degenerate if its initial solution violates
the condition that the number of allocated (occupied) cells n (C,y) ts greater than or
equal to sum of number of rows n [r) and number of columns n (c) minus one. In other
words if"n (C-) < n (r) + n {c) - \f the TP is degenerate.
It becomes difficult while computing optimal solution if the TP has degeneracy.
Therefore the degeneracy is to be first removed before going to computation of IBFS.
How to Resolve Degeneracy :
Degeneracy in TP can be resolved by allocating a minute quantity e {read as
epsilon) to an appropriate cell. This e is assumed so negligible quantity such that
a + z = a- z = a and E # 0 (also E + e = £, E - E = 0).
Where a is any quantity.
Thus zis a very minute and negligible nan zero quantity that does not alter any value
xvken added or subtracted any number of times.
e.g.: E is like one Rupee tip given to worker for loading/unloading in a business
where some lakhs or some thousands are spent in transporting the goods.Thus e does
not affect that total transportation cost.
The degeneracy may be experienced in the initial stage of optimisation or at
the subsequent iterations.
In general the degeneracy can be identified in transportation problem while
finding IBFS itself by the following points.
1. If the partial sum of availability is equal to partial sum of requirements, there
h a chance of getting degeneracy.
2. In any method (of NWCM VAM, LCEM etc.), if a row and a column are
simultaneously satisfied/exhausted and hence both deleted for next iteration,
then degeneracy will occur while optimising.

5.6,1 Degeneracy in the Initial Solution

ILLUSTRATION 1 8 - — —„____„_—_

Determine the Optimum basic feasible solution to the following Transporation

problem.
To
B Available
1 50 30 220 1
From
2 90 45 170 3
3 250 200 50 4
Required 4
[JNTU (Mech.) 99/CCC, (ECE) 99/CCC|
256 Operations Research : Theory and Practice

S o I u+i cm i
Step 1: The given Transportation problem (TP) is in the standard from (i.e,T
minimization case). Hence we can proceed directly.
Step 2 : In the given TP, the sum of availability (8) is equal to the sum of required (8).
Therefore it is balanced.
Step 3 ; Preparation of IBFS by North West Corner Method,
To
A B C_ Available
LL 1
50 30 220

From LL
90 45 170
LJL
250 200 50
Required 4 2 2
Step 4 : Optimization : Tlie number of allocated cells say (n) = 4; No. of rows 4- No. of
columns - 1 say r + c-\ = 3 + 3-1 = 5
As n < r + c - 1, there is a degeneracy. *Hence assuming e as an allocation at
the cell (1, B) such that e ^ 0 and a + E = a-E = a.
Then, the costs ut and v: are computed.
A 0 Q w. Uj + V: = C-j for allocated cells
«l + v^ = 50
LL u\ = 0 (assumed)
50 30 220 Mj +1»2 = 3 0
LJL U2 = 40 t*2 + i»l = 9 0
90 45 170 u^ + v2= 200
LL 12 = 170 -50
250 200 50
= 50 v2 = 30
(As we have five equations and six variables, we have to assume one of these as
equal to zero. Hence w.j = 0 Is assumed)
Now, calculating the value o[c^J- (u^ + V:) for non-allocated cells, to check the
optLmality, we get
A B C ^
LL | e + 340
0
50t-s i6 30 220
+250
90"* i *s 45 170
40
+ 30 2
250 200 50
170
4 2 2
50 30 -120
* The degeneracy can be detected in initial stages while finding IBFS, since the partial
sum of availability of 1st & 2nd row it., (1 + 3) is equal to partial required of A i.e., 4.
Also, when cell (2, A) is allocated with 3 units both row and column get satisfied.
Transportation Problem 257
As c-J - (u{ + v' )<0 for cell (2, B), we revise it by transfering some units given
by least figure among die corners marked by -ve on the loop i.e. e here and then the
above step is repeated.

LLJ 4-25 + 3G5

0
50 30 220

LiJ 45
+ 275
170
40
90
3
+5
LL LJ_ 195
250 200 50
Required 4 -2
50 -M5
As the values of c-' - (a- + vj)> 0 (non negative) for all non-allocated cells, the
optimal basic feasible solution is obtained.
T h e solution is,
From To Unit Cost No, of Units Total Cost
1 A 50 I 50
2 A 90 3 270
2 B 45 45e
3 & 200 2 400
4 C 50 9 100
Total 320 + 45 £

Hence the total tranportation cost is 320 units (or Rs.)

5.62 Degeneracy in the Subsequent Iterations

ILLUSTRATION 19
Optimise the following TP
Supply
120
8 5 6
80
15 10 12
80
3 9 10
150 80 50

Solution
With usual steps (as explained) the initial solution for the above TP
(minimisation 8c balanced) by North-west corner method is given below,
258 Operations Research : Theory and Practice

fltERATIOIUTABLEAUl:

/H> Supply
|l20 120
8 5 6
| 30 | 50 80
15 10 12
| 30 | 50 80
3 9 10
150 80 50
Since ihe number of occupied cells n (Cy) is 5 and No. of rows n (r) + No. of
columns i,e,, n {c) - 1 is 3 + 3 - 1 = 5, there is no degeneracy, Thus w^ and V: are
computed with usual set of rules, Also, values C^- (u^vj) for unoccupied cells are
calcaulte'd. The loop is identified for most negative.
ITERATION TABLEAU 2:
Supply
| 120 +2
8 5 6
| 30 +1 80
S2
15"-- 10 12

- LiL 10| 50 80 i=6

_vi 3
9
Demand 150 80 50 280

ITERATION TABLEAU 3 :
Supply
120 120
8 5 6
| 30 80
15 10 12
[so] 9 10
80
3
Demand 150 80 50 280

In IT - 3 we notice degeneracy* since occupied cells are only four while

n (r) + n (c) - 1 = 5 . T h u s n (Cjj) < n (r) + n {c)-\.
Therefore we have to resolve degeneracy first, by allocating e to ($2,^1) or
(S3, D 2 ) s u c h ^ a t a + e = 0 ^ E = fl and e ^ 0 and proceed with usual rules.
Note : Here degeneracy occured since both cells have turned out to be unoccupied when 30 units have
been transferred.
* The chance of oeturiiig degeneracy can be predicted in the begin nine: since partial sum
of availability of S\ &S3 (120 + 80) is equal to partial sum oT requirement of
Z)iandZ>3(150 + 50).
Transportation Problem 259
ITERATION TABLEAU 4;

D2 Supply ?iv
| 120 -9 120 0
5 Ve
+ 11 | 80 +:i
80 -4
1
15 10 ! 12
-H 1 ^® 80 -5
9 10

Demand 150 80 50 280

8 14 15

ITERATION TABLEAU
1U5:5:
D, Supply ?^
-9 +
120 0
8 5 6
+ 11 | 80 + 10 80 -4
15 10
[so +9
SO ^5
9" 10
Demand 150 80 50 230

8 14 6

ITERATION TABLEAU 6 :
Supply
| 50 120 0
8 5 6
+2 | 80 80 +5
15 10 12
| 80 +9 +9 80 -5
3 9 10
Demand 150 80 50 280
V; 8 5 6

Since C^ > (Uj + vp for all the unoccupied cells, the optimal solution is arrived.
The optimal cost calculations are shown below.
Cells From To No. of Units Units Cost Total Cost
c,, ^1 10 8 560
cl2 Si D2 E 5 5e

*1 50 6 300
260 Operations Research : Theory and Practice

80 10 800

h SO 3 240
230 - 1900 + 5e
G. Total
= 1900
Since value of £ is negligible, the total transportation cost = 1900 units.

5*7 Special Cases in TP

There are some special cases in TP and are listed below. [We may have a
combination of two or more of these specialities]
L Maximisation cases
2. Unbalanced cases
3. Multiple or alternate optimal solution case
4. Restricted transportation case
5. Conditional T.P.
6+ Trans-shipment problem.
Of the above the first two cases have been explained in the earlier illustrations
including them as first two steps oflBFS of the T.P. However, these cases are explained
again in combination with restricted and conditional transportation problems under
illustration -21 and 22 to follow. The third case has been explained in illustration 10,
11 & 12. However, the alternate optimal solution in illustration - 12 is shown here,

57.1 Multiple Optimal Solution or Alternate Optimal Solution

ILLUSTRATION 20
The optimal solution of TP given in illustration - 10 or 12 is considered here to
illustrate the alternate optimal solution.

After arriving the optimal solution by MODI or stepping stone method, we now
identify alternate optimal solution as follows
Dl D,A Z>3 D4 Supply it;

r
| 20 0
I 0
1 4
+1
3
LH-
.A*
\ 30
1
50

| 20 +4 20
V
3 4 2 5 9
Demand 20 40 30 10 100
1 2 1 0
Transportation Problem 261

With usual calculations of U: 8c u: find C{j - {ut: + u) for noil-all oca ted cells. As
all are + ve we obtain optima] solution equal to 180.

Since C- - {n- + vj) value for (OXt D%) cell is zero (or C^ - u} + vj), this eel) is as
good as any allocated cell (whose C^ = uL + vp and thus can enter the basis (allocation).
We take up this cell and construct a loop as shown in the above table. From the loop
(O^Ds) -> (O 2 , ^3) -* (0^0$ -> (0}t D2)t we transfer the units from
(O\r D^) or {O2, ^3) and the maximum possible units to transfer is only 10. Therefore
we choose the direction from (O]t D2) to {O\, Z>3). Thus the revised matrix is given
below.

[Add 10 to (O[t D3), subtract 10 to (O2, D$)f Add 10 to {O2, D2) and subtract 10
to (O^Dg), Thus {O|, D§) becomes occupied i,e,, basic while {O±t D^) becomes
unoccupied i,eM non basic]
Note ? This resembles the some Zj -Cj=Q values in sivipUx table. Comparefor your heller understanding
£>, D2 Dz Z?4 Supply u}
| go 0 Li! +4 30 0
t 2 1 4
+1 | go | 20 50 1
3 3 9 1
+3 | so +4 +9 20 0
4 2 5 9
Demand 20 40 30 10 100
2 3 0

From the above table the values of C-j - {u- + vp > 0 for all unoccupied cells,
thence it ts optimal solution.

The cost calculation are given below.

Cells From To Units Cost No. of Units Total Cost

c 1 20 20
n ^1 ^1

Cis
°\ 1 10 10

C, 2 3 20 60

o. 2 20 40

c24 °2 1 10 10

C.Y1 °* 2 20 40

Grand Total transportation cost ISO

262 Operations Research : Theory and Practice

532 Restricted TP
Here, we will consider a problem with a combination of maximisation and
restriction (or prohibition) in transportation.
In solving the problems, the rec trie ted cell is assumed to have an unaffordable
cost M such t h a t M + a-- - M - a.^ - M.

ILLUSTRATION 21
ABC agency transports material from one place to the other on commission
basis. The following are the estimated commissions per unit of materials to be
transported from the plants Pr P2 and P3 to market centres Mp M2 and Mr
Optimise the commissions to be earned by the agency (note that there is not route
available to transport from P2 to MJ. |MBA-OU-97|

Mi M2 M3 Supply
PJ 6 9 8 120
P2 - 4 2 80
P3 11 5 4 80
Demand 150 70 60 280
Solution :
As t h e given tran sport at ion p r o b l e m (TP) is in n o n standard from (since
commission is to b e maximised), t h e T P is first converted t o s t a n d a r d from by
subtracting every element from t h e greatest a m o n g all t h e elements. T h e equivalent
cost matrix is,
M| Af2 M§ Supply
120
5 2 3
80
7 9

7 80
0 6
7
Demand 150 70 GO 280
As there is no route available from P 2 to Ml? let us assign an unaffordable cost
M such that M + ai} = M - a- = M.
And the given TP is balanced as total demand = total supply = 280.
Now finding IBFS by Vogel's Approximations,
Supply Penalty
5 9 3 120 1
80 2
M 7 9
) 80
. ' 6 • 7 0
Demand 15? 70 60 280

Penally
Transportation Problem 263

Allocation to (P 3 , M} ) -> 80, Deletion

Revision for M} from 150 to 70.

M, Supply Penalty

Pi 5
L 10
2 3 50
\

80 2
M 7 9
Demand 70 60
0
Penalty M 5 6

Allocation : (PY , M{ } -> 70f Deletion

Revision : P{ from 120 to 50.

Supply Penally

50 ]
2
| 70
80 2
P2 7 9
Demand 70 60
Penalty 5 6
T

Now, we place the allocations in the equivalent cost matrix and check for
optimality. [The TP is not degenarate as the number of allocated cells (5) is equal to
no, of rows (3) phis no, of columns (3) minus 1 i.e (5)].
M,
| 70 + 1
0
5 2 3

P<>
6
M 7 9
| 80 + 10 +9
-5
0 6 7

Note : Assumed u\ - 0 and other u\ and VJ are calculated using Qj =w + vj where Cij is allocated cell
cost) and then checked for &f — (ui + VJ) > 0 for non - allocated cells.

Since Ct:' - (^ + v'•) > 0 for all non allocated cells, the solution is optimal.
264 Operations Research : Theory and Practice

Replacing the allocations in the original matrix.

Mt Supply
| 70 | 50 ISO
6 9 8
LTD, 10 80
4 9

| SO 80
11 5 4
Demand 150 70 60 280

Calculations of optimum (maximum) commission.

From To Per Unit Units Total
Commission Transported Commission
M, 6 70 420
Pi Ma 8 50 400
p-> M,> 4 70 280
A> M
z 2 10 20
p-i M, 11 80 880
Grand Total 2000
Optimum Commission = Rs, 2000.

5J.3 Conditional TP
We now illustrate a TP with maximisation, unbalanced, multiple optimal
solution and other typical conditions.
JIMISTRATIGN 22
A company has 3 factories and 3 ware houses with the following shipping cost
and other information*
Factories Sales Price Ware House
Ware House
Ff F2 F3 (Rs.) Capacity
3 9 5 34 80
W2 1 7 4 32 120
w3 5 8 3 31 150
R/Mcost 15 1& 14
Labour cost 10 9 12
Factory capacity ISO 100 130

Due to prior commitment Fs must supply 50 units ofW2 and for every unit sent to
W} an octroi ofRs> 2Aper unit is imposed, also only at F3 left out inventory costs
Re. 1/-per unit left out.
Find the optimal solution for the company so as to make maximum profits or
minimum cost/loss.
Transportation Problem 265

Solution ;
Total of warehouse capacity (t,e,t requirement here)
= 80+120+150 = 350
Total of factory capacities
= 150+100+ 130 = 380
Since X requirement * X capacities, the TP is unbalanced. Therefore we
introduce a dummy warehouse (row) whose capacity requirement is 30 (deficit) units.
In the above TP, there is sale price, raw material cost, labour cost apart from
the shipping cost. First we have to calculate the net profit or loss to each cell with the
following formula.
Profit/loss to each cell:
Cy = sales price of corresponding warehouse - shipping cost of the cell - cost
of R/M at corresponding factory - cost of labour at each factory.
An exemplary calculation for cell Cj j i .e., W^ to f\ is shown here below.
Profit /loss for Cn =34-3-15-10
= 34 - 28 = 6 (profit)
for C12 = 34 - 9 - 18 - 9 = -2 (loss) and so on.
The profit cost calculation for each cell in the dummy row will be taken as
follows,
Capacity of Dummy row == 30,
Sales price at Dummy row = 2ero (0)
Cost of R/M with F lf F2 and F 3 = 15, IS and 14 respectively.
Cost of labour with Fh F2 and F 3 = 10, 9 and 12 respectively.
Cost of shipping at each cell in Dummy row — 0
An example for dummy cell C41 i.e., from D to i^ is
C 41 = 0 - 0 - 1 5 - 10 = -25 (loss)
Sales price Requirement

wl 3 9 5 34 SO

1 7 4 32 120
5 8 3 31 150

D 0 0 0 ; 0 30
R/M cost 15 18 14 Total 380
Labour cost 10 9 12
150 100 130 Total 380
266 Operations Research : Theory and Practice

Revised TP Matrix :
F
\ *\
34-3-15-10- 3 4 - 9 - 18- 9 = 34 - 5 - - 1 4 - 80
6 -2 3
32-1-15-10= 3 2 - 7 - 18- 9 = 32 - 4 - - 1 4 - 12 = 120
6 -2 2
31-5-15-10= 31- 8 - 18- 9 = 31 - 3 - 1 4 - 12 = 150
1 -4 2

0 - 0 - 1 5 - 10= 0 - 0 - 18- 9 = 0 - 0 - 1 4 - 12 = 30
D
^25 -27 -26
Cap 150 100 130 380

Now, let us fulfill the conditions imposed.

Condition 1 :

50 units must b e transported from F{ to W2. Let us first allocate this quantity
irrespective of profitfloss due to the prior conmiitment. Here, we get a profit @ Rs. 6
per unit on transportation of these 50 units from F± to W% i-e., Rs, 300/-. And when
these 50 units are transported, further capacities at F1 a n d W2 will be readjusted as
100 a n d 70 a n d the totals will be 330. T h e profit of the Rs. 300/- we keep aside for a
while a n d a d d at the last.

Condition -2 :
Octroi @ Rs. 2 p e r unit is imposed on the units sent to Wv therefore eveiy cell
in t h e first row, a cost of Rs. 2 p e r unit is to b e deducted.
Condition -3 :
At F^ if inventory is leftout an inventory cost of Re. 1/- p e r unit is levied.
Therefore, to the cell { D, F 3 ) Le,, C 4 3 we deduct Re 1/- since, this is the only cell
representing left out inventory at F 3 .
T h u s with the above conditions, the TT matrix is re-written as follows.
F{ F2 F^ Requirement

If, (6-2) SO
4 -4 1

70 (120-50)
6 -2 2
150
1 -4 2
-27 (-26-1) 30
D -25
Capacities 100(150-50) 100 130 330 (380 - 50)
Transportation Problem 267

[Allocation made to {W2, FL) is 50 units @ Rs. 6A profit = 300.]

Now, the figures in the above TP represent the profits, which is a non-standard
case. These are to be converted to equivalent cost matrix:. This can be done by two
methods (multiply by -1 to all cells or subtract from highest value).

The equivalent matrix is given below by multiplying with (-1) to all cells,

/"l F% F$ Requirement

80
-4 4 - 1
70
-6 2 -2
150
- 1 4 -2
30
D 25 27 27
Capacities 100 100 130 330

For the above TP, we now find IBFS by North West Corner method. T h e initial
solution is given below:
F| F2 F% Requirement
| 80
80
-4 4 -1
| 20 | 50
70
W2
-6 2 -2
| 50 100 150
- 1 4 _9
30 30
D 25 27 27
Capacities 100 100 130 330

In the above TP, there is no degeneracy since n {Cl}) -n(r) + n(c)- 1

i.e., 6=4 + 3-1

We proceed to find ux and Vj values with occupied cells using Cy = u^ + Vj and

then calculate C^ - (w? + vp for unoccupied cells. A loop is constructed for cell with
most negative value of C^ - (w;- + vA>
268 Operations Research : Theory and Practice

F
2 F
$ Requirement
L SO 0
4 - 1
80 0
-4
L 20 [ 50 +2
-2
70 -2
-6 2
+3
_} 4
1 50
*_9
| 100
150 0

0 -6 30 29
D 27
25 ••-i 2 7

Capacities 100 100 130 330

v
i -4 +4 -2

h Requirement
L 80 0
4 - 1
80 0
-4
L 20 2
I 50 +2
_ 2
70 - 2
-6 i
+3
[ 20 1
]30 150 0
-1 4 -2
D
+6 [ 30 30
25 27 27
Capacities 100 100 130 330
-4 +4 -2

In the above tabie all the values C{J - (u^ - vj) > 0, therefore optimal solution is
attained.
The solution is to be transfeiTed to profit matrix by changing the sign for each
unit cost.
• As we have C^j - (u— vj) = 0 for (W^ Z7^) cell, we get an alternative optimal
solution, which can be obtained by transferring 50 units from (H^ F%) and
doing necessary adjustments.
Solution -1: Solution - II:
Req. Req.
\ 80 80 | 5Q
80
4 -4 1 4 -4 1
| 20 | 50 70 70
6 -2 2 6 -2 2
| 20 | 130 150 | 20 | 130 150
1 -4 2 +1 -4 2
| 30 30 30 30
D
-25 -27 -27 -25 -27 -27
Cap, 100 100 130 330 Cap. 100 100 130 330
Transportation Problem 269

Solution -1:

Allocated From To Unit Loss No. of Total Profit Remark

Cell or Profit Unit or Loss

Cn w
\ 4 SO 320 Profit

c21 w
'> 6 20 120 Profit

F'2 W, -2 50 -100 Loss

Ft W
?, -4 20 -80 Loss

W
?.
9 130 260 Profit

h (Dummy) .97* 30* (-810*) *Gost that can

he minimised by
not producing
Total 700-180 = 520

Solution - / / ;

Allocate From To Unit Loss No. of Total Profit Remark

or Profit Unit or Loss

cM F] w, 4 30 120 Profit

F* -4 50 -200 Loss

Ft G 70 420 Profit
F
<> -4 20 -80 Loss

F
-i w
i 2 130 260 Profit

F
(Dummy) -27* 30* (-810*) Not to produce
*
Total 800^280 = 520

The over all profit Rs. 520 can be obtained provided the production at F2 is
restricted to 70 units only instead of 100, i.e., 30 units excess produced from F9 is to
be reduced (^otherwise it will cost Rs. 810 which turns the profit of Rs. 520 to loss of
Rs, 290).

Further the profit of Rs+ 300 on allocating 50 units at (W2y F}) adds to die answer:

,\ Optimal profit = 520 + 300 = Rs. 820/-.

270 Operations Research : Theory and Practice

Practice Problems
Obtain IBFS by NWCM for the following transportation problems whose unit costs
of transportation, availabilities and requirements are given in the matrices, and
then optimize
1. [JflTU (CSE) - 2000/S] [Madras B.Sc. S I ]

Destination r> DG Supply

Source Supply
1 2 3 4
9 12 9 8 4 3
1 3 7 6 4 5
7 3 6 8 9 4
2 2 4 3 9 2
3 4 3 8 5 3 % 4 5 6 8 10 14
Demand 3 3 2 2 10 7 3 5 7 10 9

2 3 8 10 2 4

"\Coi winner Demand 3 4 5 7 6 4

Pany ^ ^ ^ A B C Supply There is degeneracy in this problem
1 6 8 4 !-l 6. [Delhi (stat) - 73]

2 4 9 3 12 5 Supply
3 1 2 6 5 2 11 10 3 7 4
Demand 6 10 15 B 1 4 7 2 1 8
C 3 9 4 8 12
3. [US -89, 0U - MBA - 92] Demand
D$ DA Supply
7.
6 4 1 5 14
Supply
S 9 2 7 16
A 11 13 17 14 250
4 3 6 2 ,,5
B 16 18 14 10 300
Demand 10 15 4
C 21 24 13 10 400
Demand 200 225 275 350 950
[IAS - 89 , AU - B.T«ch -90]

1 11 111 IV Supply
0, 13 n 15 20 200
o2 17 14 12 13 600
18 13 15 12 700
Demand 300 300 400 500
Transportation Problem 271

8. [0U-B,T«th (Mecii) - 91, JNTU - (FDB) - 92, OU - HBA -Sep. 99]

DA Supply
L9 30 50 10
70 40 60

°3 40 8 70 20

Demand 5 8 7 14 34
Answer: 1 9 x 5 + 3 0 x 2 - 3 0 x 6 +40 x 3 + 7 0 x 4 + 20 x 14= 1015
9. Find the optimum solution to the following transportation problem in which
the cells contain the transportation cost in rupees :
W, Available

h 7 6 4 5 9 40
8 5 6 7 8 30

h 6 8 9 5 20
5 7 7 8 6 10
Required 30 15 20

Answer ; 7 x 30 + 6 x 10 + 5 x 20 + 6 x 10 + 9 x 5 + 6 x 15 + 8 x 5 + 6 x 5 = 635
10. Find IBFS by matrix minimum method to die following TP, each cells value
being the unit profit, arid find the corresponding profit Also compare your
results by any other method. [0U - MBA 90]
#i £>2 &% £*4 ^5 Avail
34 55 47 24 32 250
41 28 32 46 51 150
20 31 35 47 50 175
O± 53
53 35
35 26
26 I 28
28 39
39 125
Req. 130 280 110 60 120 700
Answer: 34 x 130 + 5 5 x 120 + 28 x 150 + 31 x 10 + 35 x 110
+ 47 x 155 + 28 x 5 + 39 x 120 = 26785
11. Find IBFS to the following TP using WWC, VAM and LCEM, each cell value
being cost, and find cost. [00 - MBA - 91]
Di D9 D% D4 D7i Avail,
17 n 12 20 8 300
16 9 10 15 14 180
18 19 9 11 12 145
13 15 14 16 15 130
Req, 200 250 150 SO 120
Answer: 1 7 x 2 0 0 + 1 1 x 1 0 0 + ! x 150+ 10 x 30 + 9x 120+ 11 x 25 + 16x55 = 8385
272 Operations Research : Theory and Practice

12. Find the initial basic feasible solution to the following T.P using VAM given the
COSt matrix [OU - MBA - Sep 2001, May 92]

D D2 Supply

20 25 2S 31 300

32 28 32 41 ISO

FA 18 35 24 32 110

Demand 150 40 180 170

13. Solve the following transportation problem using matrix minimum methods
each cell value being the unit profit. [0U - MBA - Sep, 2000] *
A> Supply
4 1 3 4 4 60
9 3 2 2 3 35

3 5 2 4 4 40

Demand 22 45 18 30 135

14. A company has four factories situated in different locations and five ware houses
in different citites, The matrix of transportation cost (Rs. per unit) is given below
with capacity (unities) as factories and the requirement (units) of the ware
houses. [0U - MBA - Apr. 94]
Ware Req,
House IT III rv (Units)
A 4 8 7 6 J50
B 9 5 8 8 50
C 6 5 8 7 40
D 5 8 6 3 60
E 7 6 5 8 200
Capacity
100 80 120 100
Units
Find initial basic feasible solution by matrix method.
15, Find the initial basic feasible solution to the following transportation problem
given the unit cost matrix, using least cost entry methtjdu - MBA • Feb. 93, Sep. 99]
D, Supply
20 28 32 55 70 50

48 36 40 44 25 100

0, 35 55 22 45 48 150
Demand 100 70 50 40 80
Transportation Problem 273

16, Find the IBFS to the following transportation problem, Fiven the unit profit
Matrix, Using Vogel's approximation method, [OU - MBA - May 95]
Factory A B C Supply
i— 1 i I
10 8 8

r2 10 7 10
1] 9 7
12 14 10
Demand 10 10 8

17. Find an IBFS for the following transportation problem using- Vogel's
approximation method. Compare with the result of NWC,
[OU - MBA Apr, 93, Sep. 99, May 95]
D] D*} D% D4 Avail.

190 300 500 100 50

700 300 400 600 100

400 100 600 200 150
Req. 50 80 70 140

18+ A garment manufacturing company has two centers at Nizamabad and

Hyderabad where it manufactures special skirts of same specifications, but at
varying costs. Nizamabad centre has a capacity of 2500 skirts at cost of
production as Rs. 23/- per skirt while that at Hyderabad it can produce 2100 skirts
@ Rs. 25/- per skirt. Four boutiques are willing to purchase these skirts with the
maximum requirements as 1800, 23O0f 550 and 1750 and offer the prices as 39,
37, 40 and 36 respectively. The cost (in rupees) of the shipping one skirt from
Nizamabad center to the four boutiques is 6, 8, 11 and 9 while that from
Hyderabad is 12, 6, 8 and 5 respectively. Determine the optimal distribution
arrangement for the garment manufacturer. [iGMHi (MCA) DEC 97, JULY 99]

Hint: The above information can be formulated into a TP matrix as follows.

Boutiques Cap. Mfg. Cost
1 2 3 4
Mizamabad 39-23-6 = 37 - 2 3 - S - 4 0 - 23- 11 = 36-23 - 9 =
@ Rs. 23/- 10 6 6 4
Hyderabad 39_25- 12 = 37 - 2 5 - 6 = 40 - 2 5 - 36-25 - 5 = 25
2 6 7 6
Req. 1800 2300 550 1750

Sales Pnce 37 40 36

[Then proceed with creating a dummy row of (production) capacity 1800 and
2ero as cell cots. Notice that the given problem is non-standard, hence proceed by
assigning negative sign or subtracting from highest.]
274 Operations Research : Theory and Practice

19, ABC Enterprises is having three plants manufacturing dry-cells, located at

different locations. Production cost differs from plant to plant. There are five
sales offices of the company located in different parts of the country. The sales
prices can differ from place to place. The shipping cost from each plant to each
sales office and other related data are given in the following tables:

Production Data table :

Production Cost Maximum Capacity Plant Number

Per Unit in Number of Units
20 150 1

22 200 2

18 125 3

Shipping Cost and Demand and Sales Prices Table :

Shipping Cost:

Sales Office
Plant 1 1 1 5 9 4

Plant 2 9 7 8 3 (5

Plant 3 4 5 3 2 7

Demand and Sales Prices :

Demand 80 100 75 45 125

Sales price 30 32 31 34 29

Find the production and distribution schedule most profitable to the company.
[IGN0U * MCA - June 1997, Dec. 99)

Hint: Refer the hint of its previous problem.

20. Raju and Co is having three manufacturing centres at Hyderabad,

Vishakapatnam and Nizamabad with the production capacities as 200, 300 and
500 units respectively. Its markets are situated at Secunderabad, Vijayawada,
Tirupathi and Warangal with the demands as 250 each. The costs of
transportation are found to be Rs. 20, 40, 25 and 30 per unit from Hyderabad
to various stations respectively. From Vizag (Vishaka) these costs are 70, 40, 80
and 60 and from Nizamabad Rs. 50, 70, 30 and 40 per unit respectively. When
there is a raise in demand of the product, the markets require 50% extra. For
urgent delivery cases Nizamabad centre can produce 25% extra and can send
at Rs, 10 per unit extra load carried to various places. Hyderabad centre can
Transportation Problem 275

sub-contract 50% of its extra capacity and can supply at Rs. 20 hike in
transportation. Formulate TP and suggest the optimal policy for the case of
raised demand.
Hint: Create two more production centres at Nizaniabad and Hyderaad respectively
with 25% and 50% increased capacity and Rs, 10/- and Rs. 20/- hiked cost. Also
add increased demand to the markets. The following is the formulation,
Secunderabad Vijayawada Tirupathi Warangal Capacity
Hyderabad 20 40 25 30 200
Vishakapatnam 70 40 80 60 300
Nizamabad 50 70 30 40 500
Nizamabad-2 GO 80 40 50 125
(25% extra)
Hyderabad-2 40 60 45 50 100
(50% extra) J225
Demarvl 375 375 375

2 1 A product is manufactured by four factories, At B, C and D. The unit production

costs in them are Rs. 2, Rs. 3, Re. 1 and Rs. 5 respectively. Their production
capacities are 50, 70, 30 and 50 units respectively. These factories supply tiie
product to four stores, demands of which are 25, 35, 105 and 20 units
respectively. Unit transportations cost in rupees from each factory to each store
is given in the following table :
Stores
1 2 3 4
A 2 4 6 11
B 10 8 7 5
Factories
C 13 3 9 12
D 4 6 8 3

Determine the extent of deliveries from each of the factories to each of the
stores so that the total production and transportation cost is minimum.
[JNTU (Mech) 98]

22. A manufacturer has distribution at Agra, Allahabad and Kolkata. lliese centres
have available 40, 20 and 40 units of his product. His retail outlets require the
following number of units : A - 25, B - 10, C - 20, D - 30, E - 15,

The shipping cost per unit in rupees between each centre and outlet is given
in the table.
276 Operations Research : Theory and Practice

Retail Outlets
A B C D E
Agra 2 4 6 11 50
Allahabad 10 8 7 5 60
Kolkata 13 3 9 12 30

Determine the optimal shipping cost.

[JNTU - Mech./Prod./Chem./MechatroTHCS 2001/S]

23, A company has three plants A, B> C which supplies to ware houses 1, 2, 3T 4, 5,
Monthly plant capacities are 800, 800,900 units respectively while the monthly
requirements at the warehouses are 400, 400, 500p 400, 800 units respectively.
The unit transportation costs are shown in table below.

Determine optimum distribution for the company in order to minimize the

total transportation cost:
Warehouses
1
5 8 6 6 3
Plants 4 7 7 6 6
8 4 6 6 3

[JNTU - CSE/E.Comp E - 2001]

24* A company has three plants A, B and C and four warehouses W+ X , Y and Z.
The number of units available at the plants are 15, 25 and 5 respectively and
the demand at warehouses are 5, 15, 15 and 10 respectively. The unit
transportation costs are as follows.
W X Y Z
10 0 20 11
Plants 12 7 9 20
0 14 16 13

Find the allocation that gives minimum transportation cost

25, Find the optimal solution to following TP, given unit cost
£>} B? &$ &4 Supply
0,' 15 19 15 22 50
19 13 18 17 60
21 12 16 15 40
Demand 20 25 16 35
[ 0 U - M B A - A p r i l 94]
Transportation Problem 277

26. Find optimal solution to the following TP given unit cost of transportation
matrix.
D, D2 D3 Supply
16 20 12 200
24 8 18 160
26 24 16 90
Demand 180 120 150

[OU - HBA - Feb. 93]

27. Find optimal solution to the following TP using VAM - MODI method each cell
value being the unit cost [0U - MBA - July 2000, Dec, 2000]

D4 Supply

C 35 41 28 16 20 285

14 21 28 SO 15 145

45 18 17 29 26 165

Demand 125 125 100 100 175

28. Find optimal cost the following TP [0U - HBA 91]

Supply

12 18 13 20 50

Oc 17 11 16 15 60

11 10 14 13 40

Demand 20 25 10 35

29- Find optimal cost of the following TP [0U - MBA Nov. 94]

1 2 3 Supply

42 27 24 35 50

46 37 32 32 60

40 40 30 35 40

Demand 40 20 60 30
278 Operations Research : Theory and Practice

30. Find OBFS of following TP. Using VAM for IBFS [ 0 U - M B A - S e p . 98]

Wx W2 W.6 W{ Supply (tons)

100 120 95 110 250
85 75 100 12fl
no 105 80 100
75 85 70 90 500

f., 140 130 120 125 400

Demand 650 500 450 400

31. Consider a TP with three ware houses and four markets warehouses capacities
are : Wi = 30, W2 = 70 and Wz = 50 market demands are : Mx = 40, M2 = 30,
M 3 - 40, Af4 = 40. The unit cost movement "in Rs. is given below.

M, M2 M4 Supply
20 20 20 10 50
W2 100 80 50 40 60
70 60 60 SO 40
Demand 40 20 60 30

Find optimal solution [OU - MBA - Apt.

32, A potato chip manufarturer has three plants and four ware houses. The
shipping costs are given below. Find optimal solution after using northwest
corner rule. [0U - MBA - Apr. 99]
Plants
Warehouses 1 Avail.

5 8 6 6 235
4 7 7 6 280
8 4 6 6 110
Req. 125 160 110 230 625

33- A department store whisehs to purchase the following quantities of sarees.

Warehouses A B C D E

150 100 75 250 200

Transportation Problem 279

Tenders are submitted by 4 different manufactuers who undertake to supply

not more than the quantities mentioned below. (All types of serees combined)
Manufacturer w X Y Z
Qjy. 300 250 150 200
The store estimates that the profit per saree will vary with the manufacturer as
shown in che folowing matrix.
Sarees
A B C D E
W 275 350 425 225 150
X 300 325 450 175 100
Y 250 350 475 200 125
Z 325 275 400 250 175
How should the orders be placed [0U - HBA - May. 9Z\

5.7.4 Trans-shipment Problem

In transportation problem, the transportation or shipping occurs between
sources to destinations. And the units cannot be transferred from one source to the
other nor one destination to the other. But in practical situations, when a destination
requires more than its estimated demand, it may bring the unsold (or surplus) units
lying at another destination. This is even applicable to the source. Thus if the
transportation is, allowed among the sources as well as among destinations, apart from
the regular transportation from sources to destinations, the problem is said to be
trans-shipment. Diagrammatic ally it is shown below.

(a) Tra n s po rta I i o n pro b k m (b) Tra 11 s - s h i p m en t prob 1 cm

FIGURE 5,& : DIFFERENCE BETWEEN TRANSPORTATION AND TRANS-SHIPMENT

Thus in trans-shipment problem, sources will play the role of destinations also,
and destinations may be sources too. Thus a transportation problem of *mr sources
and V destinations will become a trans-shipment problem with (m + n) sources and
(m + n) destinations. i,e,, mxn matrix becomes (m + n) x (m + n) matrix.

And if total number of units transported is AT, then add this N at supply and
demand at each source as well as destination. Then the problem is solved as usually
with the rules of TT.
280 Operations Research : Theory and Practice

IUUSTRAT10N 23

A firm having two sources, S, and S2 wishes to ship its products to two
destinations Z>, andDr The number of units of units available atSs and S2 are 19
and 30 respectively while the demands at Dt and D2 are 25 and 15 units
respectively. The firm instead of shipping from source to destinations, decides to
investigate the possibility of trans-shipment The unit cost of transportation (in
Rs.) is given as follows:
From St to S& Dt andD2 : 3,4,5 respectively.
From S2 to / ) , andD2 : 3, S respectively
From DttoD2 :2
and the cost of upward shipment os same as that of downward shipment this
means cost ofSf andS2 is same as S2 to S} andso-on. (Symmetric)

The formulation of the above information in the form of a trans-shipment problem

is as follows ;
To Supply
From
10 + 40*
0 3 4 5
30 + 40*
s 0 3 5
40*
4 3 0 2
40*
0 5 2 0
Demand 40* 40* 25 + 40* 15 + 40* (40 + 160*)

* The total number of units available at Sj and S2 are (10 + $0 = 40) is e upl to
total demands at Dx and D2 (25 + 15 — 40), Thus maximum number of units
that can be shipped is 40 units. Since all the supply points (sources as well as
destinations) and demand points are able to supply this total number units,
these quantities are regarded as buffer stock and is added to all sources as well
as destinations at both availability and demand. Thus the TP is revised and
IBFS is then found by least cost entry method and then for OBFS, MODI
method is applied as follows :
Transportation Problem 281

D, Supply
40 +3 +1
50 0
0 3 4 5
+3 | 40
LL 70 0
3 0 3 5
+7 +6 | 40 0
40 _ i

4 3 0 2
+ 10 + 10 +4 40
5 5 9 0

Demand 40 40 65 55 200
vj 0 0 3 5
All the values of C y ' - (u- +v) > 0, the above T P has reached optimal solution.
T h e solution is SltoD2:l0 units x 5 Rs, = Rs. 50
S2 to D} : 25 units x 3 Rs, = Rs. 75
S2 to D2 : 5 units x 5 Rs. = Rs. 25
Totai = Rs. 150

However, 5 units of {D{ , Dx) cell can be transferred to cell (Du A>)- T h u s an
alternate solution that is in accordance with trans-shipment is as follows.

5, S2 Z), D2 Supply
| 40
50
0 3 4 5
30
70
3 0 3 5
[^ LL 40
4 3 0 2
| 40
D>
\ 5 | 5 2 0
Demand 40 40 65 55 200

Tlie solution to the above optimal T.P is

From To No. of Unit Cost Total Remarks

Units
Da 10 5 50 5i, 52 and Diz have each
40 units buffer while 35
s2 0| 30 3 90 units are buffer at D\
and 5 more units are
0| 5 2 10 sent from D\ to D%
Total 150
282 Operations Research : Theory and Practice

Review Questions
1. Give the applications of transportation problem in the industries.
[JNTU - C5E 98, OU - MBA M 9 1 , F 93]

2. Write the algorithm for solving a TP. IOU - MBA - M 921

3. Bring out the relation between LPP and TP. [JNTU - Mech, 97/PJ

4. T P is a case of LPP — discuss, [OU - MBA - 98, JNTU BEE si]

5. How do you formulate a TP problem as an LPP.

Explain with an example of your choice. [JNTU - EEl 92]
6. With an example of your choice, explain the algorithm of Nortli West coiner
method to obtain IBFS in a TP. [3HT0 - EEE

7. What is IBFS. Give step by step approach to find TBFS in a TP using Vogel's
approximation method. [JNTU EEE - 93 / OT]

8. Write the algorithm for finding IBFS in a T P using least cost entry method.
Give suitable example.
9. What is non-standard TP, How do you optimise in this case.
10. What is mearu by unbalanced TP. How do you find optimal solution in this case.
[FU - MBA 92, JNTU - Mech./Frod./Chem - 2001/Sl

11. What is meant by degeneracy in TP. How do you resolve it?

[JNTU - Mech 96, 98/C, CSE 94,}

12. Give step by step rules for stepping stone method for finding OBFS in a TP.
[OU - MBA - J 2000P S 2001]

13. Write the algorithm for MODI method. [JNTU - CSE 95/S]

14. Give flow chart to Find IBFS of T P using the following methods,
(a) NWCM (b) VAM (c) LCEM
15. Draw a flow chart to find OBFS of T P using
(a) Stepping stone method,
(b) MODI-method, [CSE - 92/s]

16. What is meant by a prohibited (restricted) TP, How do you proceed to find
OBFS in this case?
17. What is meant by a conditional TP. Explain with an example.
18. Write a short note on trans-shipment problem with an example.
[JNTU - Mech. 96, 97/P, 9S/C, CSE, 96, 97/S]
19. When d o you get multiple optimal solution in TP. Explain with an example,
how you identify this situation.
[Mech, 97/C 99/P,0U - MBA Dec 90, F 93, H 92, Sep. 2001]
Transportation Problem 283

20. Draw a flow chart for optimisation of TF including its initial solution.
21. When does degeneracy occur in TP?[0tl - MBA - 91 M, 94 D, 95 D, 95 M, J 2000, S 99]
22. What are the differences between general LPP and TP?
[0U * MBA - 95 Dec, $5 M, D 2000, Hay 99]

23. What is balanced TP? [OU - MBA - Apr 98]

24- What are the differences between stepping stone method and MODI Method?
tOU - MBA 93]
25. What is a TP. [OU - MBA - M 95]
26. What is shadow prices in TP I0U - MBA -98]

Objective Type Questions

1. Which of the following is not false in the case of the concept of le' introduced
to resolve degeneracy in TP,
(a) € -f- a = e (b) E - a = e
(c) e = 0 (cl) e + € = €
2. Which of the following is false in optimal T.P. whose C^ - (u^ + vfi values of
(a) all unoccupied cells must be equal to zero
(b) all unoccupied cells must be less than zero
(c) all occupied celts must be equal to zero
(d) all occupied cells must be greater than or equal to zero
3. Which of the following method may y teld same IBFS for both minimisation and
maximisation of same data-
(a) Vogel's approximation (b) least cost entry method
(c) north west corner rule (d) row minima method
4. In which of the following cases degeneracy may not appear
(a) partial sum of supplies is equal to partial sum of demands
(b) both row and column are satisfied in same step in Vogel's approximation
(c) same least cost for more than one cells
(d) no. of allocated cells is less than rows + columns - 1
5. Which of die following methods uses penalties to find IBFS
(a) no rth we s t c or n e r meth o d (b) VogeP s a p prox ima tion m e th od
(c) least cost entry method (d) column minima method
6. Which of the following method to find IBFS of transportation problem is
independent of costs /profit of transportation matrix
(a) north west corner rule (b) Vogel's approximation
(c) matrix minima method (d) row minima method
284 Operations Research : Theory and Practice
7. Multiple optimal solutions in TP are observed if
(a) Cy = u^ + Vj for some occupied cells
(b) Cjj * ui + V: for some occupied cells
(c) C- - u• + Vj for some unoccupied cells
(d) Cy ^ Uj + Vj for some unoccupied cells
8. The solution is said to be optimal in T.P if
(a) Cy > ut + ^ for all occupied cells
(b) C?y < ut- + f • for all unoccupied cells
(c) Cy > u% + v. for all unoccupied cells
(d) Cy < ^ + u; for all unoccupied cells
9. A T.P. is said to be balanced if
(a) no. of rows = no. of columns
(b) no. of allocated cells = no. of rows + no. of columns - \
(c) Total supply = total demand
(ci) no. of supply centres = no, of demand centres
10. In a restricted (or prohibited) T.F., we use M as the cost in corresponding ceil
with the concept that

11. In a maximisation case of T.P, we convert it to minimisation case by

(a) adding every cell value to highest among them
(b) subtracting every cell value from highest among them
(c) dividing cell value with lowest among them
(d) subtracting least value of each row in the corresponding row,
12- Which of the following is false with trans-shipment problem as compared to its
corresponding transportation problem.
(a) supplies can be done by demand centres
(b) demand can be adjusted among demand centres
(c) transportation is made between supply centres and demand centres
(d) a supply centre can not move its goods to other supply centre
13. In a T.P. Matrix of the order 4 x 3 , the degeneracy if found if number of
allocated cells are
(a) 11 (b)7 (c)6 (d)5
14. Anon degenerate T.P. has five allocated cells. Then tt may have
(a) 6 rows (b) 6 columns
(c) 3 rows & 2 columns (d) 2 rows & 4 columns
Transportation Problem 285

15. To find IBFS in maximising T.P, which of ihe following is wrong

(a) allocation is made at cell with highest value in LCEM
(b) allocation is made by using highest penalty at the cell with least cost
(c) all cells are multiplied by {-1) before allocating
(d) all values are subtracted from highest value among all celts.
16. If Lotal supply > total demand, then add
(a) dummy supply centre
(b) dummy demand centre
(c) dummy supply and dummy demand centre
(d) difference at any demand centre
17. Degeneracy is observed if
(a) a row and column are simultaneously filled-up in north west corner method,
(b) no. of allocated cells plus one is less than sum of no. of rows and no. of
columns in a T.P.
(c) minimum ratio is found equal for two or more basic variables in simplex
(d) any of the above
18. Given T.P. matrix as follows, cost by NWCM is

P Q R Supply
A 5 3 2 60
B 4 2 1 40
Dem. 20 30 50 100
(a) 200 (b)250 (c) 300 (d) 100
19. In the above T.P. matrix, the IBFS by LCEM gives a cost of

(a) 200 (b) 250 (c) 300 (d) none of the above
20. In which of the following method of finding IBFS, the top-left corner cell is
compulsory allocated
(a)VAM (b) ROW Minima (c) NWCM (d) LCEM
21. Which of the following is false in the case of loop drawn while optimising a T, P.
(a) every loop must have even number of corners
(b) a loop must have at least four corners
(c) closed loop may or may not be a square/rectangle
(d) the loop lines need not be horizontal and vertical lines

22. With reference to the loop draw while T.P. optimization, the minimum number
of corners is
(b)2 (c)3 (d)4
286 Operations Research : Theory and Practice

23. A loop in T.P optimisation should have at the corners,

(a) all allocated cells (b) all unoccupied cells
(c) at least one unoccupied cell (d) only one unoccupied and all other
occupied cells

24. In the case of loop drawn while optimising TP, which condition is false
(a) each row must have only one minus and one plus sign
(b) loop must be closed
(c) loop must have odd number of occupied cells at its corner
(d) shape of loop must be square

25. In IBFS of the following transportation problem, the some figures were found
missing. Identify the missed cell C- Le., cell of itht row &ndjth column

30
5 6 30
40
3 2 A 50
60
7 5 70
40 50 60

(a)C 12 (b)C 23 (d)C 3 I (d)C 3

26. The number of units missed in the above problem is

(a)'50 (b) 20 (c) 10 (d) 70

27. For the T.P. given in Q. No. 25, the transportation cost by NWCM is

(a) 410 (b)420 (c) 450 (d) 460

28. For the above problem given in Q* No. 25, the transportation cost by least cost
entry method is

(a) 410 (b) 420 (c) 460 (d) 470

29. For die problem given in Q, No, 25, which method does not develop
degeneracy at the initial stage
(a) VAM (b) LCEM (c) NWCM (d) none of these

30. The cost of transportation in a dummy cell is

(a) unity (b) zero (c) « (d) negative
Transportation Problem 287

Fill in the Blanks

L In NWCM, the allocation is independent of of the

transportation,

2. Hie top-left cell of the TP is necessarily allocated in method

3. The total supply must be equal to total demand, ill is condition is called

4, If partial sum of supply is equal to partial sum of demands dien may

appear

5, Penalties are used in method of finding IBFS in TP

6, The criterion used in LCEM is cell units of TP matrix

7, In MODI method we use ut and v- such that cost of cell (C-j) = u{ + v> for

In MODI method, if C'^- is the cost of an unoccupied cell in ih row,/' 1 column

and ut and vj are assumed costs of ft row and / ' l column, then the. optimal
solution should constitute for all unoccupied cells

9. A TP is said to be balanced if

10. A TP is said to be non-degenerate if _

11. The criterion used in VAM (to find IBFS) is_

12. If some unoccupied cells have the costs such that Cj- - {u- + t/-) = 0, then theTP
may yidd
13. In restricted TP, we use the cost 71/f to the restricted cell with the condition that

14. The epsilon (€) used to resolve degeneracy is based on the conceptual
conduions given by mathematical expression as

15. In a TP of 'IH1 rows and V columns, there exist number of basic

feasible solutions

16. A loop drawn in method of optimization of TP should consist at least

corners

17. The costs of dummy cells are taken as _ _

288 Operations Research : Theory and Practice

18. Fill up the missing numbers

| 50 T
100
4 6 2 1
60 IQ HO
3 5 7 2
R S
3 0 5 80
6
50 50 50
R= $=
r=
19. If supply exceeds demand, we add

20. AtransportatEonprobleminwhichavailablecommodityfrequentlymovesfrom
one source to another source or destination before reaching its actual
destination is called

Answers
Objective Type Questions:
Mb) 2.(c) 3. to 4,(c) 5. (b)
fi-(a) 7.(c) 8. to 9-(c> 10, (c)
11.(b) 12. id) 13. (d) 14. (d) 15. (b)
16. (b) 17-(d) 18, (b) 19, (b) 20. (c)
2L (d) 22. (d) 23. (d) 24> (d) 25.(d)
26. (c) 2r (d) 28. (d) 29. (c) 30. (b)
Fill in the Blanks :
1. cost/profit 2. northwest corner
3. rim condition 4. degeneracy
5, Vogel's approximation 6. lowest cost
7, all occupied cells 8, C'ij - im + vj) > 0 or C;j > w/ + yy
9. total supply = total demand 10. n(Ci>) = n ( r ) + - n ( 0 - 1
11. Penally 12. AJiernate or multiple optimal solution
13, M±tf.y=M 14. % ± e = a.ij ; e ?£ 0
15. m + v - 1 16. four
17. zero IS. 110, 20, 30, 50, 50
290 Operations Research : Theory and Practice

CHAPTER AT A GLANCE

Start

I
Express the problem in matrix form

Convert it into a minimisation problem,

Yes by subtracting all the values in the matrix
from the largest value in the matrix or
multiplying by (-1) to all elements

No Create dummy row/column

with deficit supply or
demand zero transportation cost

Yes

Find opportunity cost

a. Identify smallest number in each row and subtract it from every number in that row.
b. In the reduced matrix, repeat the same thing column-wise.

Make assignments one at a time (one zero in either row or column)

and eliminate row and column once the assignment is made.

Review opportunity table

a. Draw minimum possible lines on columns and/or rows such that all zeros are covered. Is the
number of
b. Choose the smallest number not covered by a lines. Subtract it from itself and every assignments equal
other uncovered number. the number of rows
c. Add this number at the intersection of any two lines and men re-assign. columns?

Optimal solution

Stop
Assignment Problem 291

6.0 Introduction
In most cases the departments will have the specialists for operating certain
critical or sophisticated machinery/equipment while in some other cases any body can
operate any machine. Whatever the type of set up may it be, it is essential for a manager
to see that maximum work is to be derived from his man power resources, which are
precious and scarce. This can be done when and only when right job is given to right
man.
Generally, though any body can do any job, all the men will not have same
efficiency and knowledge on all the jobs. The same job, one may do fast while the
other may do it slowly or a person may do one job fast and other slow. Depending on
the variations of their efficiency on individual jobs, if the jobs are assigned according
to their efficiencies, the maximum output can be derived in minimum time. Often,
the assignment of these jobs will be done based on the experience and commonsense
of the plant engineer, which may result in success in some cases but not always. The
engineer will face the hardship if it leads to any loss and becomes so difficult to escape
from the situations because there will be neither any authentic record nor a
scientific/proven methodology to safeguard him in such situations. Moreover, the
organisation may experience the irreparable loss.
The scientific and proven method to assignment is Hungarian method. This
most suitable method for job shop production (Intermittent production), is developed
by Hungarian Mathematician D. Konig. It is also known as 'Floods' technique or 'Fast
food technique' It works on the principle of reducing the cost matrix to opportunity
costs which shows the relative penalties.associated with assigning resources to an
activity as opposed to making the best or least cost assignment. If we can reduce the
cost matrix to the extent of having at least one zero in each row and column, then it
will be possible to make optimal assignment. The algorithm is as follows.

6.1 The Assignment Algorithm

Step 1 : Standard Form : Develop the cost table from the given information. The
minimisation case is supposed to be the standard form in assignment problems. T h e cost
or time or wastage etc., is included in this type. If profit, production quantity
are given they are to be maximised. In this case the given problem is to be first
converted to the minimisation form by using one of the following rules.
(a) Identify maximum cell value and subtract all the cell values from it to get
corresponding cost Values, (or)
(b) Multiply all the values with minus one (-1).
Step 2 : Balance Checking : An assignment problem (AP) is said to be balanced if the matrix is a
square i.e., the number of rows is equal to the number of columns. If the number of
rows is not equal to the number of columns then a dummy row or dummy
column must be added by putting the assignment costs for the entire dummy
cells as zeros.
Step 3 : Opportunity Cost Table : Prepare the opportunity cost table by following rules.
(a) Locate the smallest element in each row of the given cost table and then
subtract from each element of that row.
292 Operations Research : Theory and Practice

(b) In the reduced matrix obtained from 3(a), locate the smallest element in
each column and then subtract from each element of that column. Each
row and column now must have at least one zero value (This may be used as
a check).
Step 4 : Assignment : Make assignments in the opportunity cost matrix using the
following procedure.
(a) Examine rows successively until a row with exactly one unmarked zero is
obtained. Make an assignment to the single zero by making a square (D)
around it.
(b) For each zero value that becomes assigned, cross out (x) all other zeros in
the same row and/or column.
(c) Repeat steps 4(a) and 4(b) for each column also with exactly single zero
value cell that has not been eliminated.
(d) If a row/or column has two or more unmarked zeros, then choose the zero
arbitrarily.
(e) Continue this process until all zeros in rows/columns are either enclosed
by square (assigned) or crossed out by x.
Step 5 : Optimally Criterion : If the number of assigned cells is equal to the number of
rows/columns, then it is an optimal solution. The total cost (or profit in the case
of maximisation) associated with this solution is obtained by adding original
cost (or profit) figures in the occupied cells. If no optimal solution is found then
go to step - 6.
Step 6 : Marking /Draw a set of horizontal and/or vertical lines to cover all the zeros in
the revised cost table obtained from.step -5 using the following procedure.
(a) Mark (* or S ) the row(s) in which no assignment was made.
(b) Examine the marked rows. If any zeros occur in those rows, mark a tick
or star to the respective columns that contain those zeros.
(c) Examine marked columns. If any assigned zero occurs in those columns,
then tick the respective rows that contain those assigned zeros.
(d) Repeat this process until no more rows/column can be marked.
(e) Draw a straight line through each marked column and each unmarked row.
Remember :
Draw lines on 'URMC i.e., Unmarked Rows and Marked Columns. If the number of
lines drawn is equal to the number of rows or columns, then the current solution is
the optimal solution, otherwise go to step-7. (This checking of number of lines can be used
as a check point here).
Step 7 : Revising : Develop the new revised opportunity cost table.
(a) Among the cells not covered by any line, choose the smallest element say,
k, the key element.
(b) Subtract k from every element in the cell not covered by a line.
(c) Add k to every element in the cell covered by the two lines, i.e., intersection
of two lines.
(d) Elements in cells covered by one line remain unchanged.
Assignment Problem 293

Step 8 : Repeat: Repeat steps - 4 to 7 until an optimal solution is obtained.

This algorithm is shown through a flow chart as given below :
Flow chart:

Start

Express the problem in matrix form

Convert it into a minimisation problem,

Is it by subtracting all the values in the matrix
a maximisation Yes
from the largest value in the matrix or
problem? multiplying by (-1) to all elements

Create dummy row/column

No
with deficit supply or
demand zero transportation cost

Yes

Find opportunity cost

a. Identify smallest number in each row and subtract it from every number in that row.
b. In the reduced matrix, repeat the same thing column-wise.

Make assignments one at a time (one zero in either row or column)

and eliminate row and column once the assignment is made.

Review opportunity table

a. Draw minimum possible lines on columns and/or rows Is the
such that all zeros are covered. number of
b. Choose the smallest number not covered by a lines. assignments equal
Subtract it from itself and every other uncovered number. the number of rows
c. Add this number at the intersection of any two lines and columns?
then re-assign.

FIGURE 6.1 : FLOW CHART OF STEPS IN THE ASSIGNMENT METHOD (HUNGARIAN METHOD OR FLOODS TECHNIQUES)
294 Operations Research : Theory and Practice

T h e method is explained with the following illustrations.

ILLUSTRATION 1 -
Four machines, namely lathe, milling, grinding and drilling machines are to be
repaired by four maintenance operators who can perform all the jobs but differ
in their efficiency and the tasks differ in their intrinsic difficulty. The estimates
of the times each man would take to perform each is given below in the matrix:

Machine
Lathe Milling Grinding Drilling
Operator
Venkat 8 24 17 11
Shyam 13 28 4 26
Sundar 38 19 18 15
Raju 19 26 24 10
How should the repair tasks be allocated to the operators to minimise the total
man-hours.

Solution i
Step 1 : Standard Form ; As the given matrix is to minimise (man-hours) the assignment
is in the standard form. Therefore we can proceed directly.
Step 2 ; Balance Checking : The given Assignment Problem (AP) is balanced as number
of rows (4) = number of columns (4).
Step 3 : (a) Row Iteration : Select the least of each row and subtract from the
corresponding row elements.
Machine Lathe Milling Grinding Drilling
Operator
Venkat 0 16 9 3
Shyam 9 24 0 22
Sundar 23 4 3 0
Raju 9 16 14 0
(b) Column Iteration : Now, in the reduced matrix given above, select least
number of each column and subtract from its column members of row
iterated matrix.
Machine Lathe Milling Grinding Drilling
Operator
Venkat .0 12 9 3
Shyam 9 20 0 22
Sundar 23 0 3 0
Raju 9 12 14 0
Assignment Problem 295

Step 4 : Allocation : Assign the repair jobs to the men by putting the box or rectangle
( L_l) over the zeros. While allocating in this way, see that each row/column get
one and only one assignment. When a zero is enrectangled, the other zeros of
its row and column are to be crossed out. (Normally, we give preference to
allocate those zeros which are only one available in a row/column).

Machine Lathe Milling Grinding Drilling

Operator

Venkat 0 12 9 3

Shyam 9 20 0 22

Sundar 23 0 3 X
Raju 9 12 14

Thus the optimal assignment for the minimum man-hours is

Venkat to repair Lathe machine — 8 hours
Shyam to repair Grinding machine — 4 hours
Sundar to repair Milling machine — 19 hours
Raju to repair Drilling machine — 10 hours.
Total —41 hours
ILlUSTBATfGfti - 2

There arefivejobs to be assigned on 5 machines and associated cost matrix is as

follows: Machines
1 2 3 4 5
A 11 17 8 16 20
B 9 7 12 6 15
C IS 16 15 12 16
Jobs
D 21 24 17 28 26
E 14 10 12 11 15

Find the optimum assignment and associated cost using the assignment
technique.

Step 1 : Standard Form : As the given matrix is to minimise (cost), the assignment is in
the standard form. Therefore we can proceed directly.
Step 2 : Balance Checking : The given assignment problem (AP) is balanced as number
of row (5) = number of columns (5).

Step 3 : (a) Row Iteration : Select the least of each row and subtract from the
corresponding row elements.
296 Operations Research : Theory and Practice

Machines
3 4

A 3 9 0 8 12
B 3 1 6 0 9
Jobs c
1 4 3 0 4
D 4 7 0 11 9
E 4 0 2 1 5

(b) Column Iteration : Select least number of each column and subtract from
its column members.
Machines
1 2 3 4 5
A 2 9 0 8 8
B 2 1 6 0 5
Jobs
C 0 4 3 0 0
D 3 7 0 11 5
E 3 0 2 1 1

Step 4 : Allocation : Assign the jobs to the machines by putting the box or rectangle
(Gil) over the zeros. While allocating in this way, see that each row/column get
one and only one assignment. When a zero is enrectangled, the other zeros of
its row and column are to be crossed out (X) (Normally, give preference to
allocate those zeros which are only one available in a row/column).
Step 5 : Optimality Criterion : As the number of assigned cells is not equal to the number
of rows/columns, the optimal solution is not obtained. Qob D is not assigned to
any machine and machine 5 is left unassigned). Hence we move to the next step.
Machines

Jobs A

D n
E
Assignment Problem 297

Step 6 : Marking: Draw a set of horizontal and vertical lines to cover all the zeros in the
revised cost table obtained from step-5 using the following procedure.
(a) Mark ( S ) the row in which there is no assignment i.e., Row 'D'

(b) If any zeros occur in Marked row, mark the respective columns that contain
those zeros, i.e., row 7)' has zero in col. 3: Therefore mark col. 3.

(c) If any assigned zero occurs in marked column, then tick the respective
rows that contain those assigned zeros. Here marked column i.e., col. 3.
has assigned zero in first row i.e., Row A. So mark it.

(d) Repeat this process until no more rows/columns can be marked (no more).
(e) Draw a straight line through each Marked Column i.e., col. 3 and each
Unmarked Rows. i.e., col. B, C and E
(If the number of lines drawn is equal to the number of rows or columns then
the current solution is the optimal solution, otherwise go to step-7).
As the number of minimum lines connecting all the zeros (4) is less than the
number of rows/columns (5) we move to step - 7 to revise the matrix.

Step 7 : Revising: Develop the new revised opportunity cost table.

(a) Among the cells not covered by any line, choose the smallest element.
Here, This value is '2' (key element).
(b) Subtract *2' from every element in the cell not covered by a line.

(c) Add 2 to every element in the cell covered by the two lines, i.e., intersection
of two lines.

(d) Elements in cells covered by one line remain unchanged.

Then allocate again using step - 4:

Machines
1 2 3 4 5
i—i
Jobs A

B 1

E
298 Operations Research : Theory and Practice

Since all the jobs are assigned and all the machines are assigned, the optimal
solution is arrived. The optimal solution is shown in the tabular form.

Job Machine Cost

A 1 11
B 4 6
C 5 16
D 3 17
E 2 10
Total 60

ILLUSTRATION 3

Find the minimum cost assignmentfor the following problem, explaining each
step.
Workers 1 11 111 IV V

A 6 5 8 11 16
B 1 13 16 1 10
C 16 11 8 8 8
D 9 14 12 10 16
E 10 13 11 8 16

Steps 1 &2: As the given A.P has cost matrix i.e to minimise, it is the standard form. Also
as no. of row=no. of columns, given A.P is balanced.

Jobs
Workers I II III IV V
A 6 5 8 11 16
B 1 13 16 1 10
C 16 11 8 8
00.

D 9 14 12 10 16
E 10 13 11 8 16
Assignment Problem 299

Step 3 : (a) Row Iteration : Select least member of each row and subtract in its
corresponding row members.

I II III IV V

A 1 0 3 6 11

B 0 12 15 0 9

C 8 3 0 0 0

D 0 5 3 i 7

E 2 5 3 0 8

(b) Column Iteration : Select least no. of each column and subtract from
members of corresponding column.

—> Since every column is having a zero, this step results in the same matrix as
that Step 3 (a).

Step 4 : Allocation : Allocate by enrectangling ( d ) on zeros such that each row/column

will have one and only one allocation. When a zero is allocated, the other zeros
in its row and column will be crossed out. ( ^ ) . (In doing the above, preference
will be given to the zero if it is only one present in the row/column).

I II III IV V

A 1 0 3 6 1

B x 12 15 .0 9

C 8 3 0 X 0

D 0 5 3 1 7

E 2 5 3 0 '8

Step 5: Marking : As worker B is not assigned & Job V is unassigned, we move towards
optimization by marking by following rules.

(a) Mark ( ^ ) the unassigned row i.e., Row 7T

(b) Mark the columns in which unassigned marked row has zero, i.e., col. I 8c IV

(c) Mark the row in which marked columns have the assignment i.e., rows D
8c E and strike off unmarked rows and marked column by a line.
300 Operations Research : Theory and Practice

n III IV V

A •y - A--- $ - H

B 12 15 9

C -Pol-
D

All die zeros are connected with zeros and the number of lines (4) is less than the
no. of rows/columns. (5) Hence move towards step 6.

Step 6 : Optimisation: For finding optimal solution we iterate the above table as follows.
(a) Identity least digit of unlined numbers, i.e., 3.
(b) Subtract this number in all unlined numbers,
(c) Add this member at the intersected numbers (junctions of two lines).
(d) Keep all others (lined) unaltered.
Step 7: Repeat step 4 i.e., allocation
I II III IV V

The optimal assignment is

Worker Job Cost
A II 5
B I 1
C V 8
D III 12
E IV 8
Total 34
Total cost (min.) = 34 units.
Assignment Problem 301

The above assignment yields another (alternate) optimal solutions as follows.

(Refer see 6.2-2, multiple optimal solution)
A II 5
B IV 1
C V 8
D I 9
E III 11
Total 34
ILLUSTRATION 4

There are fivejobs to be assigned on each to 5 machines and the associated cost
matrix asfollows. II III IV
A 11 17 8 16 20
B 9 7 12 6 IS
Jobs
C 13 16 15 12 16
D 21 24 17 28 26
E 14 10 2 11 15

Find the optimum assignment and the associated cost using the Assignment
Technique.

Solution :
Step 1 &Step 2 : Already standard and balanced, hence proceed to step 3:
Step 3 (a): Performing row Iteration we get.
II III IV
3 9 0 8 12
3 1 6 0 9
Jobs 1 4 3 0 4
4 7 0 11 9
4 0 2 1 5

Step, 3 (b): Column Iteration :

II III IV V
2 9 0 8 8
2 1 6 0 5
0 4 3 0 0
3 7 0 11 5
3 0 2 I 1
302 Operations Research : Theory and Practice

Step 4 ; Allocation, marking and optimality check

i ii rfa iv
A 2 9 |_0j 8 8
B „.,._ ,<2 1 -^ f-Q-j- 5—
C LQ-J ^ .^. ^&. ^jg:

D 3 7 )^' 11 5
3 ©^ ^ 1— -1

r
Step 5 : Revising and improvin
I II III IV
A [o] 7 X 6 6
B 2 1 8 [o] 5
C X 4 5 X QD
D 1 5 9 3
E 3 foi 4 1 1
Solution : A -> 1 ; Total min cost = 60

6.2 Variations in Assignment Problem

The following special cases are common to occur in assignment problem.
1. Maximisation case.
2. Unbalanced AP.
3. AP with multiple optimal solutions.
4. Restricted (or prohibited) assignment.
5. Conditional assignment.
6. Crew assignment.
7. Travelling salesman problem.
8. TPasAP.
These are discussed, here below.

6.2.1 Maximisation Case in AP

In the algorithm of AP, it is already mentioned that the standard form is to
minimise the costs or time and when it is in the non-standard form i.e., to maximise
(profits or revenue), the AP is converted to standard form first before assigning.
In business practices, there may be some cases, where the profits may be
obtained by allocating the given jobs to men or machines. In such cases, the AP is said
to be a maximisation case, which is supposed as non-standard here. This matrix can
be converted into standard form i.e., minimisation by obtaining the equivalent cost
matrix. To obtain this equivalent cost matrix we choose one of the following methods
(same as those in TP).
Assignment Problem 303
1, Identify the highest value among all the cell values and subtract each value from
this highest. Put them in their respective positions.
(This means that the lost profit as compared to the highest.)
2. Multiply by minus one (-1) to all the cell values.
(This implies thai a negative profit is loss.)
After obtaining the equivalent cost matrix, we proceed as usually by the given
set of rules (algorithm) and obtain the solution. The final solution is computed with
the original.matrix i.e., profit matrix.
This is illustrated through following example.
ILLUSTRATION 5 •

A company has 4 territories and four salesmen for assignment The territories
are not equally rich in their sales potential It is estimated that a typical salesman
operating in each territory would bring thefollowing annual sales.
Territory I II III IV
Annual Sales (rs) 60000 50000 40000 30000

The four salesmen are also considered to differ in ability, it is estimated that
working under same condition their yearly sales could be proportionately as
follows.
Sales man A B C D
Proportion OJ 0.5 0.5 0.4

If criteria is to maximise expected sales, the intuitive answer is to assign the best
salesman to the richest territory and next best to second richest and so on. Verify
this answer by assignment technique.
[Dr. BRAOU - MBA 98, JNTU (Mech) - FDH-94]

Formulation :

The sales performance of each sales man in various territories can be calculated
as sample calculation shown below.

Sales man A in territory / = -r-r x 60000 = ——:—

Sales man A in territory / / = -^— x 50000 = ———

Sales man B in territory / = -rpr x 60000 = ——— and so on.

304 Operations Research : Theory and Practice

In order to provide ease in calculation let us calculate for 2.1 years instead of
1 year and consider the figures in thousands. Thus we remove three zeros in numerator
and 2.1 in the denominator. Thus the formulation of the assignment is as follows.

\Salesman A B C {)
Territory
I 42 30 30 24

II 35 25 25 20

III 28 20 20 16

IV 21 15 15 12

Standardization :

As the given figures represent the sales (per thousands of Rs. in 2.1 years) and
sales are to be maximised, it is in non-standard form. To standardise this, (i) we change
the signs of all the cell values or (ii) Subtract all values from highest. We use (ii) here
to find equivalent cost matrix.

A B C D
I 0 12 12 18

II 7 17 17 22

III 14 22 22 26

IV 21 27 27 30

Balance Checking :
In the above AP, we have four territories (rows) and four salesman (columns),
therefore it is balanced. We proceed to the next step.

Obtaining opportunity cost matrix.

(a) Row Iteration ; (b) Column Iteration :
A B C D A B C D
I 0 12 12 18 I 0 6 6 9

II . 0 10 10 15 II 0 4 4 6

III 0 8 8 12 III 0 2 2 3

IV 0 6 6 •9 IV 0 0 0 0
Assignment Problem 305

Allocation and Marking :

A B C D

I 6 (3 9 </
11 4 4 6 s
in (.) 2 9
v> s
IV - x-
Revised mairix anclrealk)catk)ii (repeating above step)
A B C D
I 4 4 7
II 2 2 4

III « ^ _ m
i ^i 1

IV
S-- ~ 0—

Re-revised matrix (repeating the above steps)

A B D A B D
c c
I m 2 9 5 I 2 2 5

II 0 GO X 9 II X X IJD 2

III 2 X 1 III 9
[o] X 1
rv 4 X X IV 4 X X 1 °l
There are two solutions possible in the above AP. The solutions are I-A, II-B,
III-C and IV-D or I-A, II-C, III-B and IV-D. The. profits will be as follows :
Territory Solution -) Solution - U[
Sales man Profit Sales man Profit
I A 42000 A 42000
=20000 = 20000
2.1 2.1
II B 25000 B 20000
11904.76 = 9523.81
2.1 2.1
Til C 20000 C 25000
= 9523.81 = 11904.76
2.1 2.1
rv D 12000
= 5714.28
D 12000
= 5714.28
2.1 2.1
Total sales in (Rs. ) 47142.82 47142.82
306 Operations Research : Theory and Practice
62.2 Unbalanced Assignment Problem
An assignment problem is said to be unbalanced if the number of rows is not
equal to the number of columns. Number of row n (r) = Number of columns n (c)
balanced AP. Number of rows n (r) ^ number of column n (c) unbalanced AP.

In other words the AP is balanced if the AP matrix is a square and unbalanced

if it is not a square matrix.

In unbalanced AP, we notice that either number of jobs are greater or lesser
than number of men or machine. In such case, a dummy row/column will be created
with zero costs to each cell. Observe the following example to convert unbalanced AP
to balanced AP.

Af, Af, Afi M9 Dummy

/l xn X
12 *13 xn x
\t *13 0

h xn x22 X
23 h X
2\ xn ^23 0

h X
31 *32 #33 h X
31 X
33 0

h *41 *42 x4Z X

4\ *42 ^43 0
Unbalanced : Machines (3) ^jobs (4) Balanced : machines (4) = jobs (4) here D is
dummy machine.

Similarly, if jobs are less, we create a dummy row with zero costs.

After converting into balanced AP, the solution can be obtained by usual steps
given in the algorithm 6.1.
It is explained through following illustrative example.
ILLUSTRATIOfti 8 — — — — -

Raju and Co has four lathe machines on which four workers operate. Any
worker can operate any machine but due to the difference in skill and machine
complexity the time of operation varies. The average times in hours when same
job done on each machine by each worker is given below:
Li L2 L4

7 6 4 9
w2 5 5 8 8
w3 4 5 4 6
w4 7 8 5 8
Assignment Problem 307

(a) Find optimal allocation,

(b) The company wants to replace the less efficient lathe with a new machine.
The probable times (in hrs) that each worker can operate is estimated as 4,5, 6
and 6 respectively. Verify whether the company has to replace any machine. If
so, which machine is to be replaced.

Here, we have two sub-problems. One is to find optimal solution for existing
machines and workers and second is to verify if any machine is to be replaced. The
first one is balanced while the second is unbalanced AP.

(a) The given problem is standard since we have to minimise the time of operation.
Also, it is balanced since it is a square matrix. The number of lathes (4) =
number of workers (4). Therefore we find the opportunity cost matrix by re y
and column iterations as shown below.
Row Iteration : Column Iteration .
L2 L4
3 2 0 5

0 0 3 3
0 1 0 9
w.
wA 2 3 0 3 w4
Now, we assign as follows. Also marking is done since H74 and L 4 are not
assigned.
J-t[ -L/2 ^$ •*-J4

1±r
wA 2

(Marking and lining) (Assignment in revised matrix)

The optimal solution is
Worker Lathe Time
4
w2 L2 5
4
308 Operations Research : Theory and Practice

L_L
Total cost j 21

(b) Now, if a new machine say N is considered for replacement. It is unbalanced,

so let us assume a dummy worker (D) with zero times on any machine which,
means that this fictitious *-:ovker allocated to any machine does not work i.e..
this machine can to be removed being less efficient.
L. N

wx 6 4 9

w9 5 .*) I? 8

4 5 4 «> o
7 8 j 8 0
D 0 0 0 0 0
No. of rows (4) * No. of col. (5)
.. we i-idci a dummy worker (D) with zero times.
NY AW we fuid the opportunity cost matrix by row iteration. [Column iteration
is no iv.rcssary siine we have zero in every column].
N

W, 3 9 0 5 0

0 0 3 3 0

0 1 0 9 9

9 3 0 1

D 0 0 0 0

Allocation

h N .

3 9
X 5

X 0 3 3 X
0 1 X 9 9

9 3 0 3 1

D X X X 0 X
Assignment Problem 309
As there is a unique zero in fourth column (5th row), Ave assign this. Obviously:
other zeros in fifth row will be crossed out which makes zero at (W2, L2) unique in the
column. Then on assigning this, zeros at (W2> Lx) and (W2, N) will be crossed oi 11. Now
zeros at (W$, Lj) and (W^t N) are unique in their columns. Assignmg them, we cross
out zeros at (W$, L3) and (Wx, L3). Obviously zero at (W±tL$) is the remaining option
to allocate. Thus we get the optimal solution as follows. [You can also procede starting
with (Wfy L3), which is unique in its row].

Optimal Solution :

Worker Lathe Time

wx New lathe 4

w2 5

w. 4

w4 5

Total 18

From the above, we conclude that L 4 is to be replaced by the new machine

which can minimise the optimal time by 3 hours i.e., total time = 1 8 hours.

6.23 Multiple Optimal Solution

There may be more than one optimal solution possible for AP. In the above
illustrations - 3 and 5, we have two solutions each. This can be identified by the zero
positions.

If we can connect all the zeros as the corners of a rectangle or a closed loop
provided the alternate corners occupy assigned, or allocated zeros, then we get
alternative optimal solution by shufling assignment from one coiner to the next.

In illustration - 3, we can connect zeros of cells (B, I), (B, IV), (E, IV), (E, III),
(D, III). By shifting the assignment rectangle by one unit along the loop we get
alternative solution. Thus the other possible solution is A - II, B - IV, C - V, D -1, E - III
get another solution shifting assignment by one cell i.e., (B, I) to (B, IV); (E, IV) to
(E, III) and (D, III) to (D, I). Similarly you can find in illustration - 5 as shown.

Yet other example is gien here below in illustration - 7

310 Operations Research : Theory and Practice

Maximisation of Unbalanced Assignment Problem :

ILLUSTRATION 7 — — —
There machinists are to be assigned for five jobs that will result in maximum
profit [JNTU (CSE) 1998]
Job
A B C D E
Machinist 1 6 8 5 10 7.5
2 7 8.5 6 7 6.5
3 5 6.5 9 8 8.5

1. Given Matrix is unbalanced. Hence we introduce two dummy rows with zero
profit.
2. Given assignment is to maximise the profit. Hence it is to be converted to
equivalent cost matrix by multiplying with (- 1) to all the cells and rewrite.
A B C D E
1 -6 -8 - 5 -10 -7.5 Multiplied by (-1) in all
2 -7 -7 the cells to convert into
-8.5 -6 -6.5
equivalent cost matrix.
3 -5 -6.5 -9 -8 8.5
*>i 0 0 0 0 0

D2 0 0 0 0 0

Row Iteration:
4 2 5 0 2.5 - 10 is subtracted
1.5 0 2.5 1.5 9 - 8.5 is subtracted
4 2.5 0 1 0.5 - 9 is subtracted
0 0 0 0 0
0 0 0 0 0
Note : Column iterations not necessaiy as every column has zero.

Allocation ;
4 2 5
0 2.5
1.5 |~o] 2.5 1.5 2
4 2.5
0 1 0.5

0 X X X X
X X X X
1 _ D — 10
2 — B — $.5
3 —C — 9
Total Profit Units 27.5
Assignment Problem 311

6.2,4 Restricted (or Prohibited) Assignment

If there is any restriction or prohibition on the assignment, of a particular cell,
it is called restricted or prohibited assignment. As we should not have to allocate in
this cell, we assume a highest penalty to this cell say M such that
M + Cy = M-C}j = M (where Q- is cost of cell in ith row and jth column). This means
that we never allocate in this cell due to its highest cost. Except for this assumption of
M to the prohibited cell, the solution method remains same. An illustration is given
below :
ILLUSTRATION 8

Five lecturers have to deal a subject common to five different branches of

engineering. Due to understanding level of the branches and the efficiency of
lecturer, the probable number of periods of 45 minutes required for each
lecturer is given in the following matrix. A particular lecturer, namely,
Mrs. Aparna refuses to go to CSE branch and CSIT students do not accept
Mr. Prasad. Obtain an optimal assignment

ECE EEE CSE CSIT MECH

Mrs* Aparna 50 50 - 20 60
Dr. Raju 70 40 20 30 40
Mr. Prasad 90 30 50 - 30
Mr. Rajesh 70 20 60 70 20
Mr. Laxman 60 50 70 90 10

Solution ;
In the above assignment matrix, it is to minimise the time i.e., in standard form.
Further it is balanced since the number of lecturers (5) is equal to the number of
branches.
But as Mrs. Aparna refuses to deal CSE and CSIT's do not accept Mr. Prasad,
these two cells will be assigned an unaffordable penalty, say M such that
=M-a = M.
Now we have the matrix as
ECE EEE CSE CSIT MECH
Mrs. Aparna 50 50 M 20 60
Dr. Raju 70 40 20 30 40
Mr. Prasad 90 30 50 M 30
Mr. Rajesh 70 20 60 70 20
Mr. Laxman 60 50 70 90 10

We now calculate opportunity cost (in terms of time) matrix by row iteration
and subsequent column iteration.
312 Operations Research : Theory and Practice

Row Iteration :
ECE EEE CSE CSIT MECH
Mrs. Aparna 30 30 M 0 40
•Dr. Raju 50 20 0 10 20
Mr. Prasad 60 0 20 M 0
Mr. Rajesh 50 0 40 50 0
Mr. Laxman' 50 40 60 so 0

Column Iteration :
ECE EEE CSE CSIT MECH
Mrs. Aparna 0 30 M 0 to
Dr. Raju 20 20 0 10 20
Mr.-Prasad. 30 0 20 M 0
Mr. Rajesh 20 0 40 50 0
Mr. Laxuian 20 . 40 60 80 0

We allocate now, with usual rules. And then marking is done.

ECE EEE CSE CSIT MECH

Mrs. Aparna 0 - 30 M A 40

Dr. Raju 20 20" - • • - © - to 20-

Mr. Prasad 30 0 20 M M

Mr. Rajesh 20 ":&•' 40 50

Mr. Laxman 20 40 00 80

Note that in the unassigned row (Rajesh), Ave have two zeros, therefore we mark
two columns. EEE and Mech and thence mark the row Prasad and Laxman which
have assignments in the marked columns.
On revising and reallocating the above matrix, we get.
ECE EEE CSE CSIT MECH

Least of iiiilinccl is 20,

added at intersetcion,
subrraded in unlied
and others unchanged
Assignment Problem 313

The above matrix yields multiple optimal solutions as follows :

Lecturer Solution I Solution II Solution III
Branch Periods Branch Periods Branch Period
Mrs. Aparna csrr 20 CSIT 20 CSIT 20
Dr. Raju CSE 20 CSE 20 CSE 20
Mr. Prasad Mech 30 EEE 30 EEE 30
Mr. Rajesh EEE 20 Mech 20 ECE 70
Mr. Laxman ECE 60 ECE 60 Mech 10
j Total 150 150 150

Total number of periods = 150 : Total time = 150 x 45 = 6750 min.

Note: Observe the rectangular loops that can be constructed with zeros loop-1 :
C
35
: C
41 C
5b 5V Tllus we
S one
alternate solution to each loop. (Loop 2 is otherwise

6.2.5 Conditional Assignment

When a condition is imposed in the assignment, it is to be satisfied before
proceeding to calculate the solution. One such problem is illustrated here below.
ILLUSTRATION B

SaiNath Institute of Science and Technology (SNIST) is providing transport to

its students in four routes. There arefour parties made their bids as given below:
Bids in routes (in '000 Rs/month)
Party
Koti Secunderabad Mehdipatnam Charminar
Pradeep travels 4 5 7 6
Raju & Co 10 5 4 4
Harika Bus 3 6 2 5
Lavanya transport 6 4 4 5
The institute wishes to allocate one route to each party. The Pradeep travels has
offered a discount ofRs. 1000/- on additional route if they are allocated more
than one routes. Find the optimal assignment to minimise their monthly costs.
Also check whether the SNIThas to consider the offer given byProdeep Travels.
314 Operations Research : Theory and Practice

Solution ":
Here, along with the regular assignment problem, we have a condition to be
checked. Let us consider the condition as fifth alternative for SNIST with the
discounted prices and create a dummy route. Now, if the discounted row is allocated
any route, we consider the offer otherwise we reject. Here, we have to remember that
the condition is implied for additional route. Thus Pradeep travels must be allocated
to at least one route and can not occupy the dummy route in first route. Thus we take
M for Pradeep at dummy.
K M Dummy
4 5 7 6 M
R 10 5 4 4 0
H 3 6 2 5 0
L 6 4 4 5 0
F (Discounted) 3 4 6 5 0
Row Iteration : Column Iteration :
K S M C D K S M C D
p 0 1 3 2 M P 0 0 1 0 M
R 10 5 4 4 0 R 10 4 2 2 0
H 3 6 2 5 0 H 3 5 0 3 0
L 6 4 4 5 0 L 6 3 2 3 0
F 3 4 6 5 0 F 3 3 4 3 0
Allocation and Marking Revised Matrix, its Allocation and Marking :
K S M K S M C D

Note: In above matrix (right side) observe the marking. First, we marked F for not having
assignment, and then marked the fifth column as the marked row has zero in this
column. This marked column (5th column) contains assigned zero at 4th row (• of L) and
hence 4th row (L) is marked. But here, this row L has one more zero at third column
(i.e., col. M) and we have to consider this also. Therefore we mark third column (M) and
subsequently the third row (H) as the assigned zero of marked column, appears at
(//, M) i.e., in row H. Thus marking is finished and then draw the lines on unmarked
rows and marked columns. As we have four lines only, we revise the matrix again. The
allocation in revised matrix is shown below.
Assignment Problem 315

K S M C D
p 0 X 2. X M

R 8 2 1 0 1

H 2 4 2 2

L 3 X X X
;
0
F X 0 2 x X
The institution has two options now, as understood from the above table.

Option -1 : Allocate routes Koti and Secunderabad to Pradeep Travels and delete Lavanya
Transport.

Option - 2 ; Allocate Pradeep Travels to Koti route and Lavanya transport to Secunderabad
route denying second route to Pradeep Travels.
Cost Calculation :

Party Option -1 Option - II

Route Cost (Rs) Route Cost (Rs)
Pradeep Travels Koti* 4000 Koti 4000
Secunderabad* 4000 _ _
Raju and Co Charminar 4000 Charminar 4000
Harika Bus Mehdipatnam 2000 Mehdipatnam 2000
LaVanya transport _ Secunderabad 4000
Total cost 14000 14000

* If Koti route is treated as second route and Secunderabad as first route to

Pradeep the cost break up will be 5000 and 3000 respectively which may not
be called an alternate solution to this problem.
I L L U S T R A T I O N 10

Engineering College has constructed the thirdfloorof a multistoreyed building.

There are five rooms in this floor meant for five classrooms. Each room has its
own advantages and disadvantages. Some have windows, some are close to
washrooms or to the canteen or good ventilation and airflow etc. The rooms are
of all different sizes and shapes. Five student leaders of different branches were
asked to rank their room preferences amongst the rooms 301, 302, 303, 304 and
305. Theirpreferences were recorded in a table as indicated below:
316 Operations Research : Theory and Practice

Si s2 S3 s4 s5
302 302 303 302 301
303 304 301 305 302
304 305 304 304 304
301 305 303
302
Most of the students did not list all the five rooms since they were not satisfied
with some of these rooms and they have left off thesefrom the list Assuming that
their preferences can be quantified, find out the best possible arrangement so
that their totalpreference ranking is minimum.
!So\uYiov\ :
The formulation for the above problem in the assignment fonn is as follows.
The figures m the matrix are the ranks given by each student leader to each
lecture hall.
s2 S4
301 - 4 2 - 1
302 1 1 5 1 9

303 9 - 1 4 -

304 3 9 3 3 3
305 - 3 4 9 -

We assign penalty M for the halls not preferred by the students.

5, 52 h *4
301 M 4 2 M 1
302 1 1 5 1 2
303 9 M 1 4 M
304 3 2 3 3 3
305 M 3 4 2 M

Row Iteration
s
\ h h h h
301 M 3 1 M 0
302 0 0 4 0 1
303 1 M 0 3 M
304 1 0 1 1 1
305 M 1 2 0 M
Assignment Problem 317

The column iteration is not required since every column has at least one zero.
Therefore we proceed to allocate.

5, h s4 h
301 M 3 1 M 0
302 0 X 4 X 1

303 1 M 0 3 M

304 1 0 1 1 1

305 M 1 9
0 M

From the above table, we have arrived the optimal solution as follows.
301 to % -302 to S,; 303 to % 304 to S2 and 305 to S4
* A typical problem, with maximisation, unbalanced and.restricted /'conditional assignment
problem is explained in the following illustrative example.
ILLO STRATSOm 11

The Indian Cricket captain Rahul Dravid has to decide firstfive batting positions
from the following five batsman whose average runs in various positions,
estimated from the past records is given below. Help him in assigning batting
order.
Position
Batsmen
II III IV
Mahendra S. Dhoni 40 40 35 25 50
Sachin Tendulkar 42 30 16 25 27
Rahul Dravid 50 48 50 60 50
V. Sehwag 20 19 20 18 25
WS. Laxman 58 60 59 55 53

If another batsman Yuvraj Singh with the following batting order positions and
corresponding record of average runs, is also considered, find the optimum
assignment (He neverplayed in the Vposition and treat that it is risky). If Yuvraj is
to be retained in the team then which of the current five batsmen will be out? What
is the change in number ofruns now?
Batting Position I II III IV V
Average runs 45 52 38 50 -
318 Operations Research : Theory and Practice

Solution :
Step 1 : Standard Form : As the given matrix is to maximise (runs) the assignment is not
in the standard form. Therefore we make it as standard form (equivalent cost
matrix) by subtracting every element from highest amongst all.
I II III IV V
M.S. Dhoni 20 20 25 35 10
Sachin Tendulkar 18 30 44 35 33
Rahul Dravid 10 12 10 0 10
V. Sehwag 40 41 40 42 35
WS Laxman 2 0 1 5 7

Step 2 : Balance Checking : The given assignment problem (AP) is balanced as number
of rows (5) = number of columns (5).
Step 3 : Preparing Opportunity Cost Matrix :

(a) Row Iteration : Select the least of each row and subtract from the
corresponding row elements.

I II III IV V
M.S. Dhoni 10 10 15 25 0
Sachin Tendulkar 0 12 26 17 15
Rahul Dravid 10 12 10 0 10
V. Sehwag 5 6 5 .7 . 0
WS Laxman 2 0 1 5 7

(b) Column Iteration : Select least number of each column and subtract from
its column members.
I II III rv V
M.S. Dhoni 10 10 14 25 0
Sachin Tendulkar 0 12 25 17 15
Rahul Dravid 10 12 9 0 10
V. Sehwag 5 6 4 7 0
WS Laxman 2 0 0 5 7

Step 4 ; Allocation : Assign the batting positions to the batsmen by putting the box or
rectangle ( EH ) over the zeros. While allocating in this way see that each
row/column get one and only one assignment. When a zero is enrectangled, the
other zeros of its row and column are to be crossed out. (Normally we give
preference to allocate those zeros which are only one available in a row/column).
Assignment Problem 319

I II III rv V
M.S. Dhoni 10 10 14 14 0
Sachin Tendulkar 0 12 25 17 15

Rahul Dravid 10 12 9 0 10

V. Sehwag 5 6 4 7 X
W S Laxman 2 0 X 5 7

Step 5 ; Optimality Criterion : As the number of assigned cells not equal to the number
of rows/columns, the optimal solutions not obtained. (V. Sehwag is not assigned
to any position and position III is left unassigned). Hence we move to marking.
Step 6 : Marking : Draw a set of horizontal and vertical lines to cover all the zeros in
the revised cost table obtained from step - 5 using the following procedure.
Mark {/) V.Sehwag's row since there is no assignment. In this row, there is a
zero in col. V and so mark col. V. Again in col. V there is assignment with first row (S.
Ganguly), Therefore mark first row.

Draw a straight line through each marked column and each unmarked row, i.e.,
on 2nd, 3rd and 5 th rows and column V.
I II III

M.S. Dhoni

Sachin Tendulkar

Rahul Dravid

V. Sehwag

W S Laxman

As the number of minimum lines connecting all the zeros (4) is less than the
number of rows/columns (5) we move to step - 7 to revise the matrix.
Step 7 : Revising : Develop the new revised opportunity cost table.

Among the cells not covered by any line, choose the smallest element, this value
is 4. Subtract 4 from every element in the cell not covered by a line. Add.4 to every
element in the cell covered by the two lines, i.e., intersection of two lines. Elements in
cells covered by one line remain unchanged.
320 Operations Research : Theory and Practice

The allocate again using step - 4.

I II Ill IV
M.S. Dhoni
Sachin Tendulkar
Rahul Dravid
V. Sehwag
W S Laxman

The number of runs is now to be read from the original matrix and the total
runs are calculated as follows :
Position Name of the Batsman No. of Runs
I Sachin Tendulkar 42
II W S Laxman 60
III V. Sehwag 20
IV Rahul Dravid 60
V M.S. Dhoni 50
TOTAL 5 Batsmen 232

Now, if Yuvraj Singh (with the averages 45, 52, 38 and 50 runs in first four
positions and never went in fifth position) is considered, the given assignment
becomes unbalanced. T o balance we create a dummy position VI in which no batsman
will make any runs (i.e., assigned zeros to all the batsmen), {step - 1}
II III IV V VI
M.S. Dhoni 40 40 35 25 50 0
Sachin Tendulkar 42 30 16 25 27 0
Rahul Dravid 50 48 50 60 50 0
V. Sehwag 20 19 20 18 25 0
WS Laxman 58 60 59 55 53 0
Yuvraj Singh 45 52 38 50 - 0

Now, the problem is converted into the standard form i.e., equivalent cost
matrix.
I n III IV V VI
M.S. Dhoni 20 20 25 35 10 60
Sachin Tendulkar 18 30 44 35 33 60
Rahul Dravid 10 12 10 0 10 60
V.~ Sehwag 40 41 40 42 35 60
WS Laxman 2 0 1 5 7 60
Ywvraj Singh 15 8 22 10 M 60
Assignment Problem 321

As it is risky to allocate the batsman Yuvraj Singh to fifth position and to avoid
this, we put an unaffordable penalty cost M to this cell in the standard form i.e., in
equivalent cost matrix.

Preparation of Opportunity Cost Matrix : Row iteration :

I II III IV V VI
M.S. Dhoni 10 10 15 25 0 50
Sachin Tendulkar 0 12 26 17 15 42
Rahul Dravid 10 12 10 0 10 60
V. Sehwag 5 6 5 7 0 25
W S Laxman 2 0 1 5 7 60
Yuvraj Singh 7 0 14 2 M 52

Column iteration
I II III IV V VI
M.S. Dhoni 10 10 14 25 0 25
Sachin Tendulkar 0 12 .25 17 15 17
Rahul Dravid 10 12 9 0 10 35
V. Sehwag 5 6 4 7 0 0
W S Laxman 2 0 0 5 7 35
Yuvraj Singh 7 0 13 2 M 27

Allocation :
I II III IV V VI
M.S. Dhoni 10 10 14 25 25
Sachin Tendulkar
0 12 25 17 15 17

Rahul Dravid 10 12 9 0 10 35
V. Sehwag 5 6 4 7 X 0
W S Laxman 2 X 0 5 7 35
Yuvraj Singh 7 13 2 M 27

As every row and every column are assigned (the total number of assignments
i.e., 6 = the number of rows or columns i.e., 6) the optimal solution is obtained. The
allocation is as follows.
322 Operations Research : Theory and Practice

Position Name of the Batsman No. of Runs

I Sachin Tendulkar 42
II Yuvraj Singh 52
III W S Laxman 59
IV Rahul Dravid 60
V M.S. Dhoni 50
Total 5 Batsmen 263

With the entry of Yuvraj Singh in second position, the other positions are
readjusted and V. Sehwag (allocated to dummy) is deleted. The increase in total runs
is 31. (i.e., 232 to 263).

Remarks :

1. In the above problem, if it so comes that Yuvraj Singh is allocated to the dummy
position then he should not be taken into the team.

2. If it is insisted that Yuvraj Singh must be included in the team in the problem
(compulsory conditional assignment) the problem can be solved by assigning
unaffordable highest penalty cost M to dummy cell against Yuvraj Singh.

6.2.6 Crew Assignment Problem

ILLUSTRATION 12 _ — _ ~ _ ^ —
An airline operates seven days a week has time table shown below. Crews must
have a minimum lay over (rest) time of 5 hours betweenflights.Obtain the pair
of flights that minimises lay over time away from home for any given pair the
crew will be based at the city that results in the smaller lay over.

Delhi-Jaipur Jaipur-Delhi
Flight No. Departure Arrival Flight No. Departure Arrival
1 ZOO AM 8.00 AM 101 8.00 AM 9.15 AM
2 8.00 AM 9.00 AM 102 830 AM 9.45 AM
3 130 P.M 2.30 P.M 103 12.00 Noon 1.15 P.M
4 6JO RM 730 P.M 104 530 P.M 6.45 P.M

For each pair, mention the town where the crews should be based ?
[JNTU-(Mech) 99/CCC, (ECE) 99/CCC]
Assignment Problem 323

.Solution :

As the service time is constant, it does not affect the stationing the crew. If all
crew members are asked to reside at Delhi (so that they start from Delhi and come
back to Delhi with minimum halt at Jaipur), then the waiting times at Jaipur for
different service time connections is shown in the following table (the figures shown
are hours). If the difference is less then 5, then it is calculated for the next day.
ITERATION TABLEAU 1 (a):
101 102 103 104
1 24 24.5 28 9.5
2 23 23.5 27 8.5
3 17.5 18 21.5 27
4 12.5 13 16.5 22

Similarly if the crew is assumed to reside at Jaipur (so that they start from Jaipur
and come back with minimum waiting time at Delhi), the table showing the waiting
times of service line connections is :
ITERATION TABLEAU Kb}:
1 2 3 4
101 21.75 22.75 28.25 9.25
102 21.25 22.25 27.75 8.75
103 17.75 18.75 24.25 5.25
104 12.25 13.25" 18.75 23.75

As the crew may be asked to either reside at Delhi or Jaipur, the minimum
waiting time from Table l(a) and l(b) can be obtained for different flight connections
by choosing the minimum value of the two corresponding waiting times, provided the
value is more than 5 hours (the desired lay over time or rest). Tliese are illustrated by
encircling in the following tableaus
Note : Thefiguresof matrix l(b) are to be transposed to compare the matrices before selecting the minimal
values.
Delhi Matrix Jaipur Matrix
101 102 103 104 101 102 103- 104
24 24.5 28 ' • " ' — ^ \

~TTo) (|T (17 ^75) 12 .25

23 23.5 27 (g2 (%2i§)(18 /7§) 13 .25
s- ^
(2L5) 27 28 .25 27 .75 24 .25 (T8^)
12.5 13 16.5 25) 75) 23 .75
324 Operations Research : Theory and Practice

ITERATION TABLEAU 2 :
101 102 103 104
1 21.75 21.25 17.75 9.5*
2 22.75 22.25 18.75 8.5*
3 17.5* 18* 21.5* 18.75
4 9.25 8.75 5.25 22*
Star (*) marked are residing at Delhi a n d others at Jaipur.
Now, applying Hungarian Method.
ROW ITERATION
101 102 103 104
12.25 11.75 8.25 0
14.25 13.75 10:25 0
0* 0.75 4 1.25.
4 1.25 4.75 16.75
Star (*) marked are residing at Delhi.
ASSIGNING & MARKING REVISION AND RE-AlLOCATiON
101 102 103 1,04 101 102 103 104

Important Note : Observe typical marking in the above table (right side).
Marking: We start with marking 4th row as it has no assignment. In the marked
(4th) row, there is a zero in 3rd (i.e., F. No. 103) column. So we mark this. We have
assignment at (1, 103) in the marked column (103) and so 1st row is marked.
Now, sine 1st row (marked) contains one more zero at (1, 104) we have to mark
co. 104. Subsequently we mark 2nd row as the marked column (104) has assigned zero
in 2nd row. With this marking is complete and then put line over unmarked row (3rd)
and marked columns (103 & 104)
RE-REVISING AND ALLOCATING
101 102 103 104 There are two solutions
%
1 0 " X Solution -1: 1 - 102, 2* - 104, 3* - 101 & 4 - 103
3 2 2
0 Solution -2: 1 - 103, 2* - 104, 3* - 101, 4 - 102

0 X 7* 12.5 Total layover =52.5 hrs.

1
Xv 0 25 The crew with 2-104 & 3 - 101 will stay at Delhi and
other at Jaipur.
Assignment Problem 325

The pattern of optimal assignment among the routes with respect to the waiting
time (in hours) is given in the following table.
Crew Residing at Solution - 1 Solution - 2
Route No. Layover Route No. Layover
1 Jaipur 1 to 102 21.25 1 to 103 17.75
2 Delhi 2 to 104 8.5 2 to 104 8.5
3 Delhi 3 to 101 17.5 3 to 101 17.5
4 Jaipur 4 to 103 5.25 4 to 102 8.75
Total Layover 52.5 52.5

62.7 Travelling Salesman Problem (or Routing Problem)

Travelling Salesman Problem : The travelling salesman problem can be solved
using assignment technique with an additional conditions on the choice of assignment,
i.e. how should a travelling salesman travel starting from his home city, visiting each
city once and returning to his home city, so that the total distance is minimum? For
e.g. given n cities with distances d-- (cost c- or time ty) from city T to city j ' the salesman
starts from cityl, then any permutation of 2, 3.... n represents the number of possible
ways for his tour. Thus there are (72—1)! possible ways of his tour and optimal value is
to be selected. From the solution of assignment, cyclic assignment is identified by
manipulating with minimum difference so as to get the solution for travelling salesman
problem.
ILLUSTRATION 13
A machine operator processes five types of items on his machine each week and
must choose sequencefor them. The set-up costper change depends on the items
presently on the machine and item to be made, according to the table.

To Item
C D
A a 4 7 3 4
B 4 a 6 3 4
From Item
C 7 6 a 7 5
D 3 3 7 a 7
E 4 4 5 7 a

If he produces each type of item once and only once each week, how should he
sequence the item on his machine in order to minimise the total set up cost
[JNTU (Mech) 99/CCC, (ECE) 99/CCC]
Solution
With usual algorithm of assignment (Hungarian method), Find the solution.
Here, the elements along principal diagonals (a's) are ignored for calculations.
326 Operations Research : Theory and Practice

ROW ITERATION COLUMN ITERATION

D A B C D
ot 1 4 0 1 (X 1 3 0 1
1 a 3 0 1 1 a 2 0 1
2 1 a 2 0 9 1 a 2 0
0 0 4 a 4 0 0 3 a 4
0 0 1 3 a 0 0 0 3 a
Now allocate as per the given set of rules.

Modify the above table by subtracting the lowest unlined element i.e. 1fromall
the elements not covered by lines and adding the same at the intersection of two lines.
D

The optimum assignment is A D, B -» A, C -H> E, D -> B and E -> C with a

minimum cost of 20.
This assignment schedule does not provide us the solution of travelling
salesman problem as it gives A^D,D^>B,B^A while B is not allowed to follow
A unless C and E are processed. This means that the assignment should be cyclic.
Therefore, now, we try to find the next best solution, which satisfies this
extra-restriction. The next minimum (non-zero) element in the matrix is 1. So we try
to bring 1 into the solution. But the element4 V occurs at two places. We shall consider
all the cases seperately until the acceptable solution is reached.
We start with making an assignment at (B, C) instead of zero assignments at (B,A).
The resulting feasible solution then will be

When an assignment is made at (C, B) instead of assignment at (C, E) the

resulting feasible solution will be
A->E,E^C,C-*B,B^>D9D^>A
The total set-up in both the programmes comes out to be 21.
Therefore both solution hold good with an extra cost of one unit.
Assignment Problem 327

6.2.8 A Transportation Problem as Assignment Problem

A transportation problem can be converted to assignment problem if the
supply/demand is in less quantity. This is illustrated through an example given below:
ILLUSTRATION 14
APSRTC can provide 6 special buses to operate for four university colleges as
required by them.
The distances in k.m., to be travelled are given in thefollowing matrix. Solve the
transportation problem as assignment problem (with Hungarian method) so as
to minimise total distance travelled.
College BKP MP KP KG Buses required
Dr. DRAOU 10 12 15 11 1
OU 8 9 7 12 2
JNTU 6 9 8 10 2
KU 6 7 6 6 1
Buses available 6

Solution :
T h e above TP is first converted into assignment model. For this we consider
each bus in each row and each column. This is done as follows.
Univ College BKP MP KP KG Required
Dr. BRAOU 10 10 12 15 15 11 1
8 8 9 7 7 12 1
OU
8 8 9 7 7 12 1
6 6 9 8 8 10 1
JNTU
6 6 9 8 8 10 . 1
KU 6 6 7 6 6 6 1
Available 1 1 1 1 1 1
Now the above assignment problem is solved by usual procedure as follows :
Since the given AP is standard (minimising distance) and balanced (row = 6 =
columns), we proceed to row/column iteration to find opportunity matrix.
ROW ITERATION
Univ. College BKP1 BKP2 MP KP1 KP2 KG
BRAOU 0 0 2 5 5 1
OUj 1 1 2 0 0 5
ou2 1 1 2 0 0 5
JNTU t 0 0 3 2 2 4
JNTU 2 0 0 3 2 2 4
KU 0 0 1 0 0 0
328 Operations Research : Theory and Practice

COLUMN ITERATION
Univ. College BKP1 BKP2 MP KP1 KP2 KG
BRAOU 0 0 1 5 5 1
ouL • • 1 • ~ 1 1 0 0 5
ou2 1 1 1 0 0 5
JNTUj 0 0 2 2 2 4
JNTU2 0 0 2 2 2 4
KU 0 0 0 0 0 0

ALLOCATION AND MARKING

Univ. College BKP I BKP2 MP KP1 KP2 KG Important Note :
BRAOU Observe marking which is
typical here. Similar
OU marking is shown in
illustration 12, under crew
assignment problem
pleaser refer.

REVISION AND RE ALLOCATION

Univ. College BKP1 BKP2 MP KP1 KP2 KG
BRAOU

JNTU2
KU
From the above table, we can derive two solutions with a total 44 km distance
and detailed calculations are shown in the following table.

Institute No. of buses Bus depot Distance (km)

Option I Option II Option I Option II
Dr. BRAOU 1 MP KG 12 11
OU 2 KP KP 14 14
JNTU 2 BKP BKP 12 12
KU 1 KG MP 6 7
Total 44 44
Assignment Problem 329
Distinction Between Assignment and Transportation Problem :

Transportation Problem is sometimes referred to as Allocation Problem. Both

these problems are the cases of linear programming problems with some distinctions.
These are distinguished as follows :

Assignment Problems Allocation or Transportation Problem

1. This problem is used to assign the jobs to machines This problem is used in transporting material
or machines to men etc. from origins (like plant) to destinations (such as
godown/market) etc.,
2. No availability (supply) and requirement Supply 8c demand are needed.
(demand) are needed to solve the problem
3. If number of rows is equal to no. of column the AP If supply is equal to demand the TP is said be
is said to be balanced otherwise unbalanced balanced otherwise unbalanced
4. There will not be any degeneracy in this problem A possibility of degeneracy either at initial stage or
subsequent stages may occur.
5. This problem is solved by converting into This can be solved without converting the given
oportunity costs (Hungarian method) costs (North west corner or Vogel's approximation
or (least cost). In VAM we use penalities in LCM
we use least cell cost in NWC we use position of
top-left in matrix.
6. T r a v e l l i n g salesmen p r o b l e m and crew Trans shipment problem is its extension.
Assignment are its extensions.
Similarties
1. Standard AP is to minimise the cost or time. Standard TP is to minimise the cost
2. Max. problem is converted to min. problem by Max. problem is converted to min. problem by
subtracting from highest among all or by subtracting from highest among all or by
multiplying all by (-1). multiplying all by (-1).
3. Restricted Assignment uses 'M?, an unaffordable Restricted TP uses 'M', an unaffordable cost.
cost.

Practice Problems
1. Find the assignment that minimises total machining time. The data is
Tasks A B CD
M/Cl
M/C2
M/C3
M/C4 ; i

[JNTU - (Mech) 92]

Answer : Mj - A, M 2 - B, M 3 - C, M 4 - D, Min time = 21 units

330 Operations Research : Theory and Practice

2. Jawaharlal Nehru Technological University has decided to give their four

departmental buildings for construction to four contractors. They have called
for tenders from approved contractors for construction of buildings. The
tender finalisation committee felt that the quotations of contractors namely
Philip & Co., Qutub & Bros., Raju & Sons and Sai Consultants are reasonable
and they decided to reject the remaining tenders. What decision the committee
should take so that amount to be spent by the university on construction is the
minimum? The following table gives quotations of four contractors in lakhs for
constructing the respective buildings.

Buildings VC Building Exams Centre Class Room Admin Building

Contractors Block

Philip & Co 10 24 30 15

Qutub &r Bros 16 22 28 12

Raju & Sons 12 20 32 10

Sai Consultants 9 26 34 16

Answer: Philip 8c Co — VC Building, Qutub 8c Bros. — Classroom block, Raju 8c

Sons — Admin. Buildings and Sai Consultants — Exams centre :•
Minimum cost is Rs. 74 lakhs.

3. A company has six maintenance groups to repair six machines. The following
table gives the return in rupees when the ith job is assigned to the/' 2 mechanic
(i,j = 1, 2, 3, 4, 5, 6). How should the jobs be assigned to the mechanics so as
to maximise the overall return?
Machine I II III IV v . • VI
Mechanic
1 9 22 58 11 19 27
2 43 78 72 50 63 48
3 41 28 91 37 45 33
4 74 42 27 49 39 32
5 36 11 57 22 25 18
6 13 56 53 31 17 28
[JNTU - (CSE) 93]
Answer: 1— VI, 2—V, 3 — III, 4—1
5 —. IV and 6— II; Maximum return = Rs. 333).
4. A department has four subordinates and four tasks to be performed. The
sub-ordinates differ in efficiency and the tasks differ in their intrinsic
difficulties. The estimate of man-hours, each man would take to perform the
task is given below :
Assignment Problem 331

I II III IV
ordinate"\
1 18 26 17 11
2 13 28 14 26
3 38 19 18 15
4 19 26 24 10

How should the tasks be allotted to men to optimise the total man-hours?
[JNTU Mech/Prodn/Chem & Mechatronics 2000,
CSE/ECE-2001]
Answer : 1 - III, 2 - 1 , 3 - I I , 4 - IV, Man hrs = 59
5. Solve the following assignment model. The assignment costs in rupees are given
below :

A
P
3
a9 R
2
S
3
T
7
B 6 1 5 6 6
C 9 4 7 10 3
D 2 5 4 2 1
E 9 6 2 4. 6
[JNTU - (Mech.) 96P/S]
Answer : A- P, B - Q C-T, D - S, E - R or
A-S,B-Q/C-T,D--P,E-R Min cost = Rs 11

6. In a textile showroom, four salesmen A, B, C and D are available to four

counters P, Qj R and S. Each salesman can handle any counter. The service (in
hour) of each counter when manned by each salesman is given below :
Counter Salesman
B C D
41 72 39 52
22 29 49 65
27 39 60 51
45 50 48 52

How should the salesman be allocated appropriate counters so as to minimise

the service time? Each salesman must handle only one counter.
(IGN0U - MCA/ADCA - Dec 99)

Answer : P - C, Q-B, R-A , S - D, Min time = 147 hrs

332 Operations Research : Theory and Practice

7. Solve the following cost minimising Assignment.

Jobs
II III IV

45 30 65 40 55
50 30 25 60 30
25 20 15 20 40
35 25 30 30 20
80 60 60 70 50

Answer: [JNTU - (Mech.) 96/P/S]

A- • / / / , B-> Vy C->/, E -^ II or
A- >III, B^ V, C^>IV, D- ^1, E^ II or min cost = 210

8. Solve the following cost minimising problems.

I II III rv V
A 2 9 2 7 1
B 6 8 7 6 1
C 4 6 5 3 1
D 4 2 7 3 1
E 5 3 9 5 1

Answer ; [JNTU - (Mech.) - 93]

A~ >II, B-> III, C->/, /)-> IV, E^ For

A -> II B -^ III C^/F D- ->/. E- > Fmin cost = 24.

9. A company has six jobs to be done on six machines; any job can be done on
any machine. The time in hours taken by the machines for the different jobs
are as given below. Assign the machines to jobs so as to minimise the total
machine hours.
Jobs
1 2 3 4 5 6
1 2 6 7 3 8 7
2 6 1 3 9 7 3
3 3 6 5 7 3 5
4 2 2 7 8 4 8
5 4 9 6 8 7 6
6 7 5 5 7 7 5
Assignment Problem 333

Answer: [JNTU - (Mech.) 91]

1 -> 4, 2 -» 2, 3-> 5, 3, 6 -» 6 or
l->4, 2^2, 3-> 5, 6, 6 -> 3 or
1 -> 4, 2 -» 2, 3 - * 5, 4->3, 1, 6 —> 6 or
Min time = 20 hours.
10. Solve the following minimal assignment problem :
Job / machine I II III IV
1 2 3 4 5
2 4 5 6 7
3 7 8 9 8
4 3 5 8 4
[JNTU - (Mech.) -92/S]

Answer : 1 -> 7, 2 -> 77, 3 -» 777, 4 -> / F o r Min : 20

1 -> //, 2 -» /, 3 -> 77, 4 -> 777
11. Solve the following minimising assignment problem :
Job/Machine 1 2 3 4 5
I 12 8 7 15 4
II 7 9 17 14 10
III 9 6 12 6 7
IV 7 6 14 6 10
V 9 6 12 10 6
[JNTU - (Mech) - 89]
Answer: I -> 3, / / -» 1,7/7 2, 7F->4, $V-> 5 or
1 -> 3, 77 -> 1, 777 -> 4, 7F-> 2, V-> 5 Min : 32

12. Minimise the following assignment problem :

(1) (2) (3) (4) (5) (6)
1 9 22 58 11 19 27
2 43 78 72 50 63 48
3 41 28 91 37 45 33
4 74 42 27 49 39 32
5 36 11 57 22 25 18
6 | 3 56 53 31 17 28 |

[JNTU - (Mech.) 90]

Answer : 1 -> (4), 2 -» (1), 3 -» (6), 4 ^ (3), 5 -> (2), 6 -> (5) or
1 -> (4), 2 -> (6j, 3 -> (2), 4 -> (3), 5 -» (5), 6 -> (1) Min cost = 142.
334 Operations Research : Theory and Practice

13. Four salesmen are to be assigned to four districts. Estimates of the sales revenue
is hundred of Rs. for each salesmen are as under :

Districts
Salesmen A B C D
1 320 350 400 280
2 400 250 300 220
3 420 270 340 300
4 250 390 410 350
Give the assignment pattern that maximises the sales revenue. [IGNOU -1991]

Answer : 1 -> C, 2 -> A, 3 -> D, 4" -» B Max Rs. .1490/-

14. Find the assignment that minimises the total time in hours.

A B C D E
Ml
M2
M3
Answer : My - E, M 2 - D, M$ -A, Min time = 8

15. The owner of a small machine shop has four machinists available to assign jobs
for the day. Five jobs are offered with expected profit for each machinist on
each job are as follows.
Jobs
A B C D E
I 60 76 48 99 80
II 69 82 59 71 57
Machinists
III 85 90 109 69 79
IV 46 62 85 75 78
Which jobs is to be declined with determining the optimum schedule ?
[JNTU - (Mech.) 95]

Answer : I-DJI-B, III - C, IV - E Dummy -A (decline); Profit : 368

16. The owner of a small machine shop has four machinists available to assign to
jobs for the day. Five jobs are offered with expected profit for each machinist
on each job as follows :
A B C D E
1 62 78 50 101 82
2 71 84 61 73 59
3 87 92 111 71 81
4 48 64 87 77 80
Assignment Problem 335
Find by using the assignment method, the assignment of machinists to jobs that
will result in a maximum profit. Which job should be declined.
[JNTU - (ECE/Mech) - 93/CCC]

Answer: 1^>D,2^>B, 3 -> C, 4 -> E : Max. profit = Rs. 376.

and Job A is declined
17. In a machine shop, a supervisor wishes to assign five jobs among six machines.
Any one of the jobs can be processed completely by an)' one of the machines as
given below :

Machine
A B C D ' E F
Job 1 13 13 16 23 19 9
2 11 19 26 16 17 18
3 12 11 4 9 6 10
4 7 15 9 14 14 13
5 9 13 12 8 14 11
The assignment of jobs to machines be on a one-to-one basis. Assign the jobs
to machines so that the total cost is minimum. Find the minimum total cost.
[JNTU - (Mech.) 95]

Answer : 1 -» F, 2 -> A, 3 -» E, 4 -+ C, 5 -> D Min cost = 43.

18. A department head has six jobs and five subordinates. The subordinates differ
in their efficiency and the tasks differ in their intrinsic difficulty. The
department head estimates the time each man would take to perform each task
as given in the effectiveness matrix below :

Task
A B C D E F
Man 1 20 15 26 40 32 12
2 15 32 46 26 28 20
3 11 15 2 12 6 14
4 8 24 12 22 22 20
5 12 20 18 •10 22 15

Only one task can be assigned to one man. Determine how should the jobs are
allocated so as to minimise the total man hours. Find the minimum total man
hours. [JNTU - {Mech.) 88]

Answer : 1 -> F, 2 -> A, 3 -» £, 4 -> C, 5 -> D, or

1 -> B, 2 -* F, 3 -> C, 4 -^ A, 5 -> D min time = 55 hours.
336 Operations Research : Theory and Practice

19. A truck company on a particular day has 4 trucks for sending material to 6
terminals. The cost of sending material from some destination to different
trucks will be different as given by the cost matrix below. Find the assignment
of 4 trucks to 4 terminals out of six at the minimum cost.
Trucks
A B C D E
Terminals 1 3 6 2 6 5
2 7 1 4 4 7
3 3 8 5 8 3
4 6 4 3 7 4
5 5 2 4 3 2
6 5 7 6 2 5
[JNTU - 95/S]
Answer : 1 -> C, 2 , 3 -»/l, 6 -> Z), Min cost = 8.
20. Solve the following unbalanced assignment problem of minimising total time
for doing all the jobs.
Jobs
1 2 3 4 5
Operator 1

[JNTU - (CSE) 95, (CA) May 85]

Answer: 1 -> 4, 2 -» 1,-3 -> dummy i.e., 6, 4 -> 5, 5 -> 2, 6 -» 3

Min time = 1 6 units.
21. A company has 4 machines of which to do 3 jobs. Each job can be assigned to
one only and only one machine. The cost of each job on each machine is given
in the following table :
Machine
W X Y Z
Job A 18 24 28 32
B 8 13 17 18
C 10 15 19 22
What are job assignments which will minimise the cost? [JNTU 96/P]
Answer: A -» W9 B ->X, C -> Y, or A ->W,B^Y,C->X,Z* = 50.
Assignment Problem 337

22- Five operators have to be assigned to five machines. The assignment costs are
given in the table below.
Machine
I II III IV V
Operator A 5 5 - 2 6
B 7 4 9 3 4
C 9 3 5 - 3
D 7 2 6 7 2
E 6 5 7 9 1
Operator ^4 cannot operate machine III and operator C cannot operate machine
IV. Find the optimal assignment schedule. (JNTU - (Mech.) 88]

Answer : A - IV, B - III, C - II, D-I, E- For

A-IV, B -III, C-V,D-IIf E-I, min cost = 15
23. Four new machines M^M%?M<$ andM 4 to be installed in a machine shop.
There are five vacant places, A, B, C, DandE available. Because of limited
waiting space machine M2 cannot be placed at C and M 3 cost (L of machine i
to place j in rupees, is shown below.

A B C D E
My 9 11 15 10 11

M2 12 9 - 10 9

M3 - 11 14 11 7
M4 14 8 12 7 8

Obtain the optimal solution using best starting solution. [JNTU - (Hech.) 88/S]
Answer: Ml-A9M2-BtMs-E9M4-D; cost = 32
24. The secretary of a school is taking bids on the four school bus routes. Four
companies have made the bids as detailed in the following table.
Bids
Company Route 1 Route 2 Route 3 Route 4
1 4000 5000 - -
2 - 4000 4000
3 3000 - 2000 -
4 - - 4000 5000

Suppose each bidder can be assigned only one route. Use the assignment model
to minimise the school's cost at running the four bus routes.
338 Operations Research : Theory and Practice

Hint: Since some of the companies have not made job for certain routes, assign
a very high bids M for all such routes. Then apply assignment algorithm.
Answer : Company 1 -^ route 1, Company 2 -> route 2,
Company 3 —> route 3, Company 4 —> route 4
Minimum cost = Rs. (4000 + 4000 + 2000 + 5000 = Rs. 15000.

25. Five workmen have to be assigned to repair five machines. The assignment
costs are given in the following table below :
Machines Lathe Milling Jig Boring Shearing SPM
Workmen

Ramu 5 5 2 6
Kishore 7 4 2 3 4
Prasad 9 3 5 - 3
Sandeep 7 2 6 7 2
Pradeep 6 5 7 9 1

Ramu cannot repair Jig boring machine and Prasad cannot repair Shearing
machine. Find the optimal assignment schedule. If Pradeep, the specialist in
SPM repairs is exclusively used for SPM and will not be sent to repair any other
machine what will be the effect on the above assignment? In case Pradeep
refuses to repair SPM, do you find any effect on the minimum repair cost?
Answer : Solution set - 1: Ramu - Shearing, Kishore - Jig boring,
Prasad - Milling, Sandeep - Lathe, Pradeep - SPM or
Solution set - 2 : Ramu - Shearing, Kishore - Jig boring,
Prasad - SPM, Sandeep - Milling and Pradeep - Lathe.
Minimum cost = 1 5 .
If Pradeep is exclusively kept for repair of SPM then the solution will be unique
i.e., the solution set 1. There will be no effect on the minimum cost if Pradeep
refuses to repair SPM since we can use the second solution" set.

26. Four operators Oh O2, O3 and O4 available to a manager has to get four jobs
Ji*J2>J$ a n c U4 done by assigning one job to each operator, given the times
needed by different operators for different jobs in the matrix below.

h h h H
Ox 12 10 10 8
o2 14 12 15 11
6 10 16 4
8 10 9 7
Assignment Problem 339

(i) How should the manager assign the jobs so that the total time needed for
all four jobs is minimum?
(ii) If]ob/ 2 is not to be assigned to operator 0 2 , what should be the assignment
and how much additional total time will be required?
[JNTU - (CSE) 95/S]

m m
Answer: (i) Oj —>j§, 0 2 ~>J2> ^ 3 ~*J4> ®4 ~~*Ji> time =
34
m m m e =
(ii) Oj ->J2> O2 ~^J4> ^ 3 ~*Jl> ®4 ~~*fa ^ ^6
.-. Additional time required = 36 - 34 = 2 units of time.
27. Five swimmers are eligible to complete in a relay team which is to consist of
four swimmers, swimming four different swimming styles; back stroke, breast
stroke, free style and butterfly. The time taken for the five swimmers — Anitha,
Bhavana, Chandi, Dolly and Easwari — to cover a distance of 100 meters in
various swimming styles are given below in minutes, seconds.
Anitha swims the back stroke in 1 : 09, the breast stroke in 1:15 and has never
competed in the free style or butterfly.
Bhavana is a free style specialist averaging 1:01 for the 100 meters but can also
swim the breast stroke in 1:16 and butterfly in 1:20.
Chandi swims all styles — back 1:10 butterfly 1:12 free style 1:05 and breast
stroke 1:20.
Dolly swims only the butterfly 1:11 while Easwari swims the back stroke 1:20
the breast stroke in 1:16 and the free style 1:06 and the butterfly 1:10. Which
swimmer should be assigned to which swimming style? Who will not be in the
relay.
Hint: The assignment matrix with time expressed in seconds and adding a
dummy style to balance it is given by
Back Stroke Breast Free Style Butterfly Dummy
Stroke
Anitha 69 75 - - 0
Bhavana - 76 61 80 0
Chandi 70 80 65 72 0
Dolly - - - 71 0
Easwari 80 76 66 70 0
Answer; Anitha will be in breast stroke (times 75 sees).
Bhavana will be in free stroke (time 61 sees).
Chandi will be in back stroke (time 70 sees)
Dolly will not participate (dummy).
Easwari will be in butterfly (time 70 sees).
Dolly will be out of the relay.
Total minimum time in the relay = 276 sees or 4 min 36 sees.
340 Operations Research : Theory and Practice

28 Alpha corporation has four plants each of which can manufacture any one of
four products: Production costs differ from one plant to another as do sales
revenue. Given the revenue and cost data below, obtain which product each
plant should produce to maximise profit.
Sales revenue (Rs. '000) product Production costs (Rs. '000)
Plant 1 2 3 4 Plant 1 9 3 4
A 50 68 49 62 A 49 60 45 61
B 60 70 51 74 B 55 63 45 69
C 55 67 53 70 C 52 62 49 68
D 58 65 54 69 D 55 64 48 66

[JNTU, 2002 (Set-2) & (Set-4)]

Hint: Construct the profit matrix by using the fact : Profit matrix = revenue
matrix - cost matrix.
To make use of minimisation technique subtract each element of profit
matrix from the maximum element which will be 8. Then apply
assignment rule in usual manner.
Answer : A - 2 > B - 4, C - 1, D - 3; Max. Profit = Rs. 22000/-

29. A small aeroplane company, operating seven days a week serves three cities
A, B and C according to the schedule shown in the following table. The lay over
cost per stop is roughly proportional to the square of the layover time. How
should planes be assigned the flights so as to minimise the total lay over cost?

Flight No. and Index From Departure To Arrival

AXB A 09 AM B Noon

A,2B A 10AM B 01 PM

A$B A 03 PM B 06 PM
A4C A 08 PM C Mid Night

AbC A 10 PM C 02 AM
BVA B 04 PM A 07 AM

B2A B 11AM A 02 PM

B%A B 03 AM A 06 PM

CXA C 07 AM A 11AM
C2A C 03 AM A 07 PM

(Bombay DMS 83; Maduiai BE (Mech) 78)

Assignment Problem 341

Answer : Min Layover = 1066 units

Flight No. 1 2 3 4 5

Departure route AXB A2B A^B CXA C2A

Arrival route B%A BYA B2A A4C

30. A trip from Hyderabad to Vijayawada takes 6 hours by bus. A typical time table
of bus services in both directions is given below.

Hyderabad - Vijayawada Vijayawada - Hyderabad

Route No : Departure Arrival Route No. Departure Arrival

a 06.00 12.00 1 05.30 11.30

b 07.30 13.30 2 9.00 15.00

c 11.30 17.30 3 15.00 21.00

d 19.00 01.00 4 18.30 00.30

e 00.30 06.30 5 00.00 06.00

The cost of providing this service by the transport company depends upon the
time spent by the bus crew (driver and conductor) away from their places in
addition to service times. There are five crews. There is a constraint that every
crew should be provided with 4 hours of rest before return trip again and should
not wait for more than 24 hours for the return trip. The company has residential
facilities for the crew at Hyderabad as well as at Vijayawada. Obtain the pairing
of routes so as to minimise the cost.
Answer:

Crew 1 2 3 4 5

Residence Hyderabad Vijaywada Vijaywada Hyderabad Vijaywada

Service No. d\ 2e 3a 64 5c

Waiting time 4.5 9.5 9.0 5.0 5.5

(hr)

Minimum total waiting time = 33.30 hours.

31. An Airline that operates seven days a week, has the time table shown below
crews must have a minimum layover of 1 hour between flights. Obtain the
pairing of flights that minimises lay over time away from home. For any given
pairing, the crew will be based at the city that returns in the smaller layover for
each pair. Also mention the town where crew should be based.
342 Operations Research : Theory and Practice

Hyderabad - Tirupati Tirupati - Hyderabad

Flight No. Depart Arrive Flight No. Depart Arrive
1 7:30 AM 9:00 AM 2 7:00 AM 10:00 AM
3 8:45 AM 9:45 AM 4 7:45 AM 10:45 AM
5 2:00 PM 3:30 PM 6 11:00 AM 2:00 PM
7 5:45 PM 7:15 PM 8 6:00 PM 9:00 PM
9 7:00 PM 8:30 PM 10 7:30 PM 10:30 PM
[JNTU - (Mech.) 95/CCC]
Answer: 1 -> 10, 3 -> 6, 5 -> 8, 7 -> 2, 9 -> 4
Minimum layover time 28:75 hours..

32. An airline that operates seven days a week, has the time table shown below.
Crews must have a minimum lay over of 6 hours between flights. Obtain the
pairing of flights that minimises lay over time away from home. For any given
pairing the crew will be based at the city that results in the smaller layover. For
each pair also mention the town where crew should be based.

Delhi - Kolkatta Kolkata - Delhi

Flight No. Depart Arrive Flight No. Depart Amve
1 7:00 AM 9:00 AM 101 9:00 AM 11:00 AM
2 9:00 AM 11:00 AM 102 10:00 AM 12:00 Noon
3 1:30 PM 3:30 PM 103 3:30 PM 5:30 PM
4 7:30 PM 9:30 PM 104 8:00 PM 10:00 PM

(ICWA June 1990)

Answer : 1 —> 103 (base at Delhi) 3 -^ 101 (base at Delhi)

2-> 104 (base at Delhi) 4 -» 102 (base at Kolkatta).
33. Solve the following travelling 'salesman problem' given by the following data :
C12 = 16, C 13 = 4, C 14 = 12, C 23 = 6, C 34 = 5 , C25 = 8, 3 3 5 = 6 and C 45 = 20
Where C- = C^and C{j= <* ifi=j
Answer : l - ^ 2 ^ 5 ^ 3 ^ 4 - ^ & cost = Rs. 47

34. A medical representative has to visit five stations^, Bf C, D and E. He does not
want to visit any station twice before completing his tour of all the stations and
wishes to return to the starting station. Costs of going from one station to
another are given below. Determine the optimal route.
Assignment Problem 343

A B C D E
A OO 2 4 7 1

B 5 OO 2 8 2

C 7 6 OO 4 6

D 1C 3 5 OO 4
E 1 2 2 4 CO

^4^JB->C^Z)^£->/l. [JNTU - (Mech.) 96/P]

35. Solve the travelling salesman problem in the matrix shown below :
To
1 2 3 4 5
From 1 oo 6 12 6 4

2 6 OO 10 5 4

3 8 7 oo 11 3

4 5 4 11 DO 5

5 5 2 7 8 oo

Answer : ; Z = 27. [JNTU- (Mech.)96/S]

36. A sales man has to visit five cities A, B, C, D and E. The distance (in hundred
miles) between the five cities are as follows :

From • A B C D E

A - 7 6 8 4

B •7 - 8 5 6

C 6 8 - 9 7

D 8 5 9 - 8

E 4 6 7 8 -

If the salesman starts from city A and has to come back to city A, which route
should he select so that the total distance travelled is minimum.
[IGNOU - (MBA) Assignment 96, JNTU - 2003/S (Set - 4)]

Answer : min distance = 30 hundred miles.

344 Operations Research : Theory and Practice

37, A machine operator processes five types of items on his machine each week,
and must choose a sequence for them. The main costs involved are setup costs
and priority. The setup costs per change depends on tlie item presently on the
machine and the setup to be made, according to the following table :
To item
A D
From item A a 4 7 3 4
B 4 a 6 3 4
C 7 6 a 7 5
D 3 3 7 a 7
E 4 4 5 7 a

In addition, there is a priority rating among the items, A and B having a higher
priority than C, Z), E. This may be interpreted as adding an additional cost of
5 when any high-priority item immediately follows a low-priority item. If he
processes each type of item once and only once each week, how should be
sequence the items on his machine? [JNTU - (Mech.) 96/P, OU - MBA 2001/S]

Hint: See illustration - 13

Additional Problems
1. Find the optimal solution to the following assignment problem, each cell value
being the processing time in minutes.
h
20 15 14 19 30
24 18 31 28 19
32 21 25 20 27
D 34 20 12 16 19
31 26 25 21 22

2. Find the optimal assignment cost.

F G H I
A
Assignment Problem 345
3. Solve the following assignment problem given the cost matrix.
Workers /i h h
24 27 18 20
26 23 20 31
24 22 34 26
19 21 21 22
30 25 28 27
[0U - MBA 93]
4. The quotations (Rs. in lakhs) received for 4 projects from four contractors are
given below :
Projects C4
c2 c3
A 7 3 7 8
B 5 7 4 7
C 9 5 8 5
D 10 8 9 2
If only one project is to be awarded to one contractor, find out the assignment
of projects to contraction in order to minimise the total cost. [OU - MBA Apr. 94]
5. Solve the following assginment problem given the profit matrix.
J\ Jz h J4 Jb
12 6 9 6 10
6 12 11 10 10
10 11 10 8 4
6 4 9 8 11
12 13 8 6 5
[0U - MBA Dec. 1995]
6. A national truck rental service has a surplus of one truck in each of the cities 1,
% 3, 4, 5 and 6 and a deficit of one truck in each of the cities 7, 8, 9, 10, 11 and
12. The distances in km between the cities with a surplus and the cities with a
deficit are displayed below :
7 8 9 10 11 12
31 62 29 42 15 41
12 19 39 55 71 40
17 29 50 41 22 22
35 40 38 42 27 33
19 30 29 16 2 23
72 30 30 50 41 20
How should the truck be dispersed so as to minimise the total distance travelled?
[0U - MBA - May 92, March 99, July 2000]
Answer: (1—>11,2—>8, 3 - > 7 , 4—> 9,.5 —> 10, 6 -» 12, Distance = 125km)
346 Operations Research : Theory and Practice

7. Check if the following assignment is optimal. If not find the optimal.

Jobs
7i h h h h
Men 20 15 14 - 30
24 - 31 28 19

32 211 - 20 27
MA - 20 12 16 19
31 26 25 21 -
[0U - MBA Dec. 2000]
8. The following cost matrix related to an assignment problem
Machines
1 2 3 4 5
7 3 4 6 2
5 2 8 7 5
Problem
6 9 6 5 2
10 8 3 9 4
(i) Find the optimum assignment
(ii) If man 2 should not be assigned to machine 2, what is the best assignment
[0U - MBA Sep. 1998]

9. Fire workers are available to work with the machines and the repsective costs
(in Rs.) associated with each worker's machine assignment is given below. A
sixth machine is available to replace one of the existing and the associated costs
and also given below.
(i) Determine whether the new machine can be accepted.
(ii) Determine also optimal assignment and the associated saving cost.
Machines
Mx M2 M3 M4 M5 MQ

wx 12 3 6 - 5 9
w2 4 11 - 5 - 8
Workers w 8 2 10 9 7 5
s
w4 - 7 8 6 12 10
5 8 9 4 6 -
Assignment Problem 347

10. Solve trje following assignment problem given the cost matrix.
D D2 Z>4
\
6 9 5 2 6

°t 2 5 8 7 7
o* 7 8 6 9 8

o4 6 2 3 3 5

o. 9 3 8 9 7
o* 4 7 4 6 8

11. A project work consists of four major jobs for which four contraction have
submitted tenders. The tender amounts quoted in thousands of rupees are
given in the matrix below.
Jobs

Jx h h 74
15 29 35 20
c2 21 27 33 17
Contractors
c3 17 25 37 15

c4 14 •31 39 21

Find the assignment which minimises total cost of the projects each.contractor
has to be assigned one job only. [OU - MBA March -1999] .

12. Suggest optimum assignment of four workers^ B, C and D to four jobs I, II,
III, IV the time taken by different workers in completing the different jobs is
given below.
Jobs
II III IV.
8 10 12 16
11 11 15 8
Workers
9 6 5 14
15 14 9 7

The indicate the total time taken is completing the job.

[OU - MBA Aug/Sep 1999]
348 Operations Research : Theory and Practice

13. A company is facing with the problem of assigning four different salesmen to
four territories for promoting its sales. Territories are not equally rich in their
sales potential and the salesman also differ in their ability to promote sales, the
following tables gives the expected annual sales (in thousands of Rs.) for each
salesman if assigned to various territories. Find the assignment of salesmen so
as to maximize the annual sales.
1 9 3 4

1 60 50 40 30
2 40 30 20 15
3 40 20 35 10
4 30 30 25 20
[JNTU C.S.E 2003/S (Set - 2)]

Answer : Sl->Tl (60), S2 -> T2 (30), S3 -» Ts (35), S4 -» T4 (20) or

S^ ^T2 (50), S2 -> Tx (40), Ss -» T3 (35)and 5 4 -* T4 (20), Max. Profit = 145000

14. There as five jobs to be assigned, one each to 5 machines and the associated
cost matrix is as follows :
I II III IV V
J°b ^ \
A 11 17 8 16 20
B 9 7 12 6 15
C 13 16 15 12 16
D 21 24 17 28 26
E 14 10 12 11 15
[AGRA M.Sc. MATHS. 1990, M.S. BAR0DA. B.Sc, 78, MADRAS IIT (M.TECH) 78,
DELHI B.Sc. (MATHS) 77, MBA 74, M.Sc. MATHS 71, INDIAN STATISTICAL INSTITUTE 71]

Answer : A -> /, B -> IV, C -> V, D -» ///, E -> / / Min. cost = 60

15. Solve the following minimal assignment problem.

I II III IV V
Job ^^^
A 1 3 2 3 6
B 2 4 3 1 5
C 5 6 3 4 6
D 3 1 4 2 2
E 1 5 6 5 4
[MEERUT STAT 1974, MATHS 92]

Answer : (A -> /, £-> IV, C -> / / /

Assignment Problem 349
16. Solve the following minimal assignment problem.
I II III IV
lob ^"-\
A 12 30 21 15
B 18 33 9 31
C 44 25 24 21
D 23 30 28 14
[MEERUT B.Sc. 1990/S]
Answer : (A -> /, B -> ///, C -> //, D -> IV)
17. Solve the following minimal assignment problem.
I II III IV
Tob
A 2 3 4 5
B 4 5 6 7
C 7 8 9 8
D 3 5 8 4
[MEERUT 1975, 91, Delhi 68]
Answer: -> /, 3 -> //, C -> III £•-» IV, or A ->/, C -> IF, D -> ///)
18.
Location I II III IV V
Project
A 15 21 6 4 9
3 40 21 10 7
9 6 5 8 10
D 14 8 6 9 3
E 21 18 18 7 4
[AGRA M.Sc. MATHS 1986, GAUHATI STAT. 69]

Answer : (A /, C •

19. Solve the assignment problem represented by the following matrix which gives
the distances from the customers A, B, C, D, E to the depots, a, b, c, d and e
each depot has one car. How should the cars be assigned to the customers as
to minimize the distance travelled?

A 160 130 175 190 200

B" 135 120 130 160 175
C 140 110 155 170 185
D 50 50 80 80 110
E 55 35 70 80 105
[MEERUT MATHS 1980, STAT. 87, CALCUTTA (APP. MATHS) M.Sc. 76,
KURUKSHETRA. M.Sc. MATHS 75, R0HITKHAND 82]
Answer: Total distance = 570 km (A -> e, B -» c, C-+b, D -^ a, E-->d)
350 Operations Research : Theory and Practice

20. The assignment cost of assigning any one operator to any one machine is given
in the following table.

Operators I II III IV
Machine
A 10 5 13 15
B 3 9 18 3
C 10 7 3 2
D 5 11 9 7

Find the optimal assignment by Hungarian method [JNTU Mech. 2003/S (Set - 4)1
Answer: A -> // (5), B -> IV (3), C -> /// (3) and D -> / (5), Total cost = 16 units

21. Solve the following assignment problem

Jobs 1 2 3 4
Operators
A 10 12 9 11
B 5 10 7 8
C 12 14 13 11
D 8 15 11 9

[MADRAS B.Sc. MATHS 84, MADURAI B.Sc. (MATHS) 83, MADURAI M.E. 1984]
Answer: (A -> / / / , B -> /, C -~> II', D ^ IV) min cost = 37

22. Solve the following assignment problem

Jobs I II III IV V
Machines

A 11 .10 18 5 9
B 14 13 12 19 6
C 5 3 4 2 4
D 15 18 17 9 12
E 10 11 19 6 14

[DELHI B.Sc. (MATHS) 93, 89]

Answer : (A -^ II, B -> V, C-> ///, D-^ IV, E ^ I) and min cost = 39
Assignment Problem 351

23. A team of 5 horses and 5 riders has entered a jumping show contest. The
number of penalty points to be expected when each rider rides any horse in
shown below.
Riders R2 Rs R4
Horses
Hi 5 3 4 7 1
H2 2 3 7 6 5
H3 4 1 5 2 4
H4 6 8 1 2 3
Hn 4 2 5 7 1
How should the horses be alloted to the riders so as to minimize the expected
loss of the team [MEERUT M.Sc. (MATHS) 84]
Answer: (Hl -> R5, R //4 //5
* oss = m m

24. Solve the following assignment problem. The elements given in the matrix are
the profits in Rs. derived for such assignment.
Machines P Q R S
Jobs
A 51 53 54 50
B 47 50 48 50
C 49 50 60 61
D 63 64 60 60
Answer : A -> R (54), B -> £ ( 5 0 ) , C -^ S (61), D -> Q (63) Max. Profit = Rs. 228
[JNTU C.S.E 2003/S (Set - 2)]
25. Find the optimal solution for the assignment problem with the following cost
matrix.
I II III IV
A 5 3 1 8
B 7 9 2 6
C 6 4 5 7
D 5 7 7 6
Answer : (A -> ///, B -» IV, C -> //, D -» /) min cost = 16 [DELHI 71]

26. Solve the following assignment problems.

Men I II III IV V
Machines
A 1 3 2 8 8
B 2 4 3 1 5
C 5 6 3 4 6
D 3 1 4 2 2
E 1 5 6 5 4
(A -> /, /F, C - / / / , Z) -> //, £ -> F) [MEERUT STAT. 74]
352 Operations Research : Theory and Practice

27. Solve the following assignment problems

I II HI IV
Tasks ^^^^-^
A 5 12 19 11
B 5 10 7 8
C 12 14 13 11
D 8 15 11 9
Answer : (A~>I, B -> ///, C -H> II, D -> IV) min cost = 35
28. Solve the following assignment problems
-^Persons I II III IV V

A 12 8 7 15 4
B 7 9 17 14 10
C 9 6 12 6 7
D 7 6 14 6 10
E 9 6 12 10 6
[MADURA! B.Sc. (MATHS) 89, OSMANIA M.Sc. (STAT.) 84, IAS (MATHS) 89, BOMBAY B.Sc.
(STAT) 85, KERALA B.Sc. (ENGG) 85, CALCUTTA B.Sc. (MATHS) 84]
Answer : (1) (A -» ///, £ -> /, C -» //, D -> IV, E -» V)
(2) {A -> ///, £ -> /, C->IV, / ) - > / / , £ -> F)
29. Solve the following assignment problems
\^Tobs I II III IV V
Machines^^-^
A 8 4 2 6 1
B 0 9 5 5 4
C 3 8 9 2 6
D 4 3 1 0 3
E 9 •5 8 9 5
[MADURAI B.Sc. (MATHS) 89, CALCUTTA B.Sc. MATHS 84]

Answer: (A -* V, B -> / , C • V, D -> ///, E -> //) min cost = 9

30. The jobs A, B, C are to be assigned to three machines X, Y, Z. The processing

costs (Rs) are as given in the matrix shown below. Find the allocation which will
minimize the overall processing cost.
-----^Machines X Y Z
Jobs
A 19 28 31
B 11 17 16
C 12 15 13
Answer : {A -> X, B -> Y; C -> Z) [BOMBAY (DIP OPER. MAN) 74]
Assignment Problem 353

31. A project work consists of four major jobs for which four contractors have
submitted tenders. The tender amounts quoted in lakhs of rupees are given in
the matrix below.
a b c d
Contractor^^^
1 10 24 30 15
2 16 22 28 12
3 12 20 32 10
4 9 26 34 16

Find the assignment which minimizes the total cost of project [each contract
has to be assigned at least one job] [MEERUT 1980, ISI (Dip) 76]

Answer: Three alternative assignments are

4 ^ a)
>, 3-xJ, 4->aj
(3) (1 -> c, d, 3 -» b, 4 -> a) min cost = Rs. 71,00,000
32. A company producing a single product and selling it through five agencies
situated in different cities. All of a sudden there is a demand for the production
at another five cities not having any agency of the company. The company faced
with the problem of deciding as how to assign the existing agencies to dispatch
the product to needy cities in such a way that the total travelling distance is
minimised. The distance between the surplus and deficit cities (in kilometers)
is given by
\ . Deficit
^ \ city A' B' a D' E'
SurplusX.
city \ ^
A 10 5 9 18 11
B 13 19 6 12 14
C 3 2 4 4 5
D 18 9 12 17 15

E 11 6 14 19 10

Determine the optimum assignment schedule

[MADRAS (M.Tech.) 77, MEERUT (STAT) 76]

Answer: {A -> A\ B^> C, C -» D', D^>B\ E -> £') optimal distance = 39 km

354 Operations Research : Theory and Practice

33. Find the minimum cost solution for the 5 x 5 assignment problem whose cost
coefficients are as given below.

1 2 3 4 5
-2 -4 -8 -6 -1
0 -9 -5 -5 -4
-3 -8 0 -2 -6
-4 -3 - 1 0 -3
-9 -5 -9 -9 -5
[ROORKEE M.E. (Mech.) 77/IIT KARAGPUR (DIP) 78]
Answer: (1 -> 3,. 2 -> 2, 3 -> 5, 4 -> 4, 5 -> 1) or
(1 -» 4, 2 -> 2, 3 -» 3, 4 -» 5, 5 -> 1 optimal cost = 36)

34. Six wagons are available at six stations A, B, C, D, E and F. these are required
at stations, I, II, III, IV, V and VI. T h e mileage between various stations is given
by the following table.

I II III IV V VI
A 20 23 18 10 16 20
B 50 20 17 16 15 11
C 60 30 40 55 8 7
D 6 7 10 20 100 9
E 18 19 28 17 60 70
F 9 10 20 30 40 55

How should the wagons be transported in order to minimize the total mileage
covered [ALIGRAH (C0MP. Sc) 77]

Answer: (A -> IV, B -> VI, C -» V, D^ III, E->I,F^>II) total mileage = 66

Review Questions
1. Write a short note on assignment problem.

2. Give the step by step procedure for solving assignment problem by Hungarian
method.

3. Discuss how you get multiple solutions in an assignment problem.

4. What is meant by unbalanced assignment problem.

[OUT - MBA 94, 0U - MBA Dec. 2000, OU - MBA Apr. 94,
JNTU Mech/Prod/Mechantronics/Chem. - 2001]
Assignment Problem 355

5. Explain in detail assignment model of LP problem. How does it differ from

transportation problem. [JNTU Mech. 96]

6. Distinguish between assignment and allocation problem, [JNTUCSE-98]

7. Write a short note on travelling salesman problem. [OU -MBA Sep. 98, Dec. 2000

JNTU CSE - 97, Mech. 99, Mech/Prod/Mechatronics/Chem. 2001/S]

8. Distinguish between assignment and travelling salesman problem.

9. What is routing problem? How does it differ from assignment problem.

10. How can you convert a transportation problem into an assignment? Explain
with an example.

11. Discuss the maximisation case in assignment problem.

12. What is meant by restricted (or prohibited) assignment? Explain how you can
solve the AP in such case.

13. How do you apply assignment in the case of allocating crew in transport
agencies/railways/airways etc.

14. Discuss the applications of assignment problems.

15. Give the solution method by Flood's technique to assignment problem in the
form of flow chart.

16. What are the difference between transportation prolem and an assignmen
problem? [OU - MBA Feb. 1993, Dec. 1990, March/April 1999]

Objective Type Questions

1. Optimality is said to be reached in AP if

(a) C-j > it} + Vj for occupied cells

(b) Z - Cj is positive for all cells

(c) highest time of middle jobs > least time of first or last series

(d) no. of minimum lines connecting all the zeros of opportunity matrix is at
least equal to its order
356 Operations Research : Theory and Practice

2. An assignment problem is said to be balanced if

(a) no. of rows = no. of columns
(b) no. of allocated cells = no. of rows + no. of columns - 1
(c) total supply = total demand
(d) no. of zero in the cost matrix is equal to the order

3. In a prohibited AP, a penalty cost 'AT is used such that

(a) a»±M = aj (b) M ± ai} = a-
{c)M±ai;j = M (d)Af±a iy =0

4. In. a maximisation case of A.P. we convert it to minimisation case by

(a) adding every cell value to highest among them
(b) subtracting every cell value from highest among them
(c) dividing every cell value with lowest among them
(d) subtracting least value of each row in the corresponding row
5. If number of rows exceeds the number of columns by 2, then we add
(a) a dummy row (b) 2 dummy rows
(c) a dummy column (d) 2 dummy columns

6. The cost of assignment in a dummy cell is

(a) unity (b) zero
(c) °o (d) negative

7. An opportunity cost table of AP is prepared by

(a) subtracting every element from highest among them
(b) multiplying by (-!) in all cells
(c) subtracting least of each row in corresponding row and then least of every
column in corresponding column
(d) transposing the matrix
8. Wliile revising the opportunity cost table of AP, we put the lines across
(a) marked rows and unmarked columns
(b) marked rows and marked columns
(c) unmarked rows and marked columns
(d) unmarked rows and unmarked columns
Assignment Problem 357

9. which of the following is not be marked while revising the AP,

(a) a row which has no assignment
(b) a column which has no assignment
(c) a row in which marked column has an assigned zero
(d) a column in which marked row has an assigned zero

10. While revising the opportunity cost table of AP, least among unlined number
i& taken to
(a) add at unlined numbers
(b) subtract at double lined numbers
(c) add at single lined numbers
(d) subtract at unlined numbers
11. If the number of minimum lines is less the order of matrix then it means
(a) optimisation is already reached
(b) one more line is to be drawn
(c) allocation is wrong
(d) there is a scope for further optimisation

12. In AP, we get multiple optimal solutions if

(a) a loop can be constructed with zeros having assignment at alternate corners
(b) minimum number of lines is greater than the order
(c) minimum number of lines is equal to the order
(d) minimum number of lines is less than the order
13. In a travelling salesman problem the additional condition on the choice of
assignment is to find
(a) in what direction should the salesman travel from home city
(b) cycle in which he has to travel
(c) a solution such that he has to return to the home city
(d) any of the above
14. A matrix of travelling salesman problem contains
(a) no figures along at least one row
(b) no figures along at least one column
(c) no figures along the diagonals
(d) no figures along the principal diagonal
358 Operations Research : Theory and Practice

15. An opportunity cost matrix of AP should have

(a) at least one zero in each row
(b) at least one zero in each column
(c) both (a) & (b) of the above
(d) no zeros in at least one column or row

16. In an unbalanced AP of 5 x 4 matrix, while finding the opportunity cost matrix,

we need not perform
(a) row iteration (b) column iteration
(c) total matrix iteration (d) marking 4 x 4 matrix
17. In an optimal solution of A.r. of 4 x 4 matrix the minimum number of lines we
must get at least is
(a) 3 (b)15 (c)4 (d)7

18. The main criterion used in Hungarian method to solve an assignment problem
is to calculate
(a) operating cost (b) maintenance cost
(c) opportunity cost (d) over heads costs

19. In an assignment whose cost matrix is given by

I II III
A 5 10 15
B 6 12 6
C 11 7 7

The optimal assignment is

(b) , 4 - > / , £ - > /7
(d) none of the above

20. The total cost in the above AP is

(a) 18 (b)20
(c)17 (d) none of the above

21. An assignment problem (AP) can be converted as transportation problem (TP)

by assuming supply/ demand to each row/column as
(a) zero (b) unity
(c) infinity (d) no relation
Assignment Problem 359

22. An assignment problem can be solved by applying Johnson Sc Jackson's rules

of sequencing if conditions is imposed.
(a) unity supply/demand (b) zero supply/demand
(c) no passing (d) infinite supply/demand

23. Men are allocated to various routes so as to result in minimum idle time or cost
and decisions is made regarding where these men have to be asked to reside.
This is performed by method
(a) transportation problem (b) travelling salesman problem
(c) crew assignment (d) routing or network problem

24. Which of the following is always a balanced problem

(a) transportation (b) assignment
(c) travelling salesman (d) none of the above

25. The column iteration is not necessary while preparing opportunity matrix in
an assignment if
(a) every column contains at least one zero after row iteration
(b) an unbalanced assignment of order m x n, where m > n
(c) every column has at least one zero initially
(d) any of the above.

Fill in the Blanks

1. If an assignment problem is square it is called

2. We can convert a maximisation AP to minimisation AP by multiplying every

cell with

3. Costs of dummy row/column are usually taken as

4. We put lines across rows and column while

revising opportunity cost table of AP

5. The optimisation is said to be reached in AP if minimum number of lines

covering all zeros is equal to

6. If a rectangular loop constructed by all zeros with alternative assignment on

the corners of the loop can be obtained, the AP will have

7. If an AP has condition that the solution must be cyclic, then the AP is said to
be
360 Operations Research : Theory and Practice

8. The optimal solution of AP of n xn order should at least get

minimum number of lines connecting all zeros in the opportunity matrix
9. In a restricted AP with elements x-, we use *Af as penalty with a condition that

10. While revising the opportunity cost matrix of AP, we identify least among
unlined numbers and to unlined numbers and " to double
lined (intersecting) numbers.

Answers
Objective Type Questions :
l.(d) 2-(a) 3. (c) 4. (b) 5. (d)
6.(b) 7.(c) 8.(c) 9.(b) 10. (d)
11-(d) 12. (a) 13. (d) 14. (d) 15. (c)
16. (a) 17. (c) 18. (c) 19. (b) 20. (a)
21-(b) 22. (c) 23. (c) 24. (c) 25. (d)
Fill in the Blanks :
1. balanced AP 2.-1
3. zeros 4. unmarked, marked
5. no. of rows or columns or order of matrix 6. alternate or multiple optimal solution
7. travelling salesman problem 8.M

9.M±xij = M 10. subtract, add

362 Operations Research : Theory and Practice

CHAPTER AT A GLANCE

Without production run With production run

(Listantaneous supply) (Continuous supply)
Inventory Management 363

7.0 Introduction
It was assumed in the past that accumulation of adequate stocks was necessary
and beneficial for the manufacturing concerns but, today large stocks are viewed with
alarm and referred to as "graveyard" of a business and the need for control is felt. In
any manufacturing concern the cost of material constitute 40% to 60% of total cost of
finished product and 26% to 30% of material cost are supposed as inventory carrying
cost. Therefore, large inventory can eat the profits. Thus there is more pressure for
maintaining liquidity in trading and industry. But on the other hand, holding of
inventories is needed in order to avoid keeping men and machines idle for the want
of materials. Also, the marketing demands a wide variety and availability at any time.
In this complex situation, the inventory management is really a challenging job.
Moreover, inventory structure varies very widely from industry to industry, depending
on nature and type of product, production technology, collaborating country, nature
and extent of the imported materials, production cycle etc.

7.1 Classification of Inventory

Different authors have classified the inventory in different ways. These are as
follows :
1. Based on Specific Purpose of Holding :
(a) Lot Size Inventory : These are held for two reasons
1. The large discounts on large quantities or lots.
2. The material is available in lots easily rather than exact amounts.
(b) Fluctuation Inventory : These inventories exist because of fluctuation in
demand or supply. Buffer or safety stocks are of this kind.
(c) Anticipation Inventory : These are built up in advance for a big selling
season, a promotion program or a plant shut down period, e.g.,
production and stocking of Diwali crackers, coolers, etc.
(d) Transportation Inventory ; These, also called transit or pipeline
inventories arise due to transportation of inventory items to distribution
centres and customers from various production centres. The amounts of
transportation inventory depend on the time consumed in transportation
and the nature of demand.
(e) De-coupling Inventory : If various products stages operate successively,
then the event of breakdown of one or any disturbance at some stage can
affect the entire system. This kind of inter dependence is not only costly
but also disruptive. To reduce this inter-dependence these inventories are
maintained.
2. Based on the Quantities and Measurement:
(a) Single or Piece Wise : e.g. pencil, chocolates.
(b) Unit Wise : e.g. set of (6) tea cups, dinner set etc.
(c) By Volume : e.g. oils, gases etc.
(d) By Weight: e.g. food grains, grease etc.
364 Operations Research : Theory and Practice

3. The Most Common Way of Classification of Inventories :

(a) Production Inventories : Items, which are used to make products, raw
materials, and bought out finished components (BOF) and bought out
semi finished components (BOSF).
(b) Maintenance, Repair and Operating (MRO) Inventories : Items which do
not form a part of final product but are consumed indirectly in the
production process like spare parts, consumable items, etc.
(c) In Process Inventories : Semi-finished products at various stage of
production.
(d) Finished Goods Inventories : completed product ready for dispatch.
(e) Miscellaneous Inventories : which arise out of the above four type of
inventories such as scrap, surplus and obsolete items, which are to be
disposed off.

7.2 The Operating Doctrine

In an inventory control situation, two basic decisions are to be taken which are
usually referred to as operating Doctrine of Inventory. These are :
(i) How much to order?
(ii) When to order?
To answer these questions one should be aware of the costs involved and other
factors those influence these decisions.

7.3 The Inventory Costs

Various costs associated with inventory control are often classified as follows.
(i) Set Up Cost: The cost associated with setting up of the machinery before
starting the production, and is independent of the quantity ordered for
production.
(ii) Ordering Cost: This is a cost associated with ordering of raw material for
production. Advertisement, consumption of stationery and postage,
telephone charges, rent for space used by purchase department, travelling
expenditure, etc., constitute the ordering costs.
(Hi) Purchase or Production Costs : This is cost of purchasing or producing a
unit of item. This is very important when discounts are allowed.
(iv) Carrying or Holding Cost: The costs associated with the storage costs like
rent, interest on the money locked up, insurance of stored equipment,
production, taxes, depreciation of equipment etc.
(v) Shortage or Stock Out Cost: The penalty cost for running out of stock (i.e.,
when item cannot be supplied on the customer's demand). This cost
includes the loss of potential profit through sales of items and loss of
goodwill, in terms of permanent loss of customers and its associated lost
profit in future sales.
(vi) Salvage Cost or Selling Price : When the demand for certain commodity
is affected by quality stocked, decision problem based on a profit
maximization criterion includes the revenue from selling.
Inventory Management 365

7.4 Terminology Used in Inventory

1. Demand : Demand is the number of units required per period and may be
either known exactly (or deterministic) or known in terms of probabilities
(probabilistic or stochastic) or completely unknown.
2. Lead Time : The time gap between placing of an order and its actual arrival
in the inventory is termed as lead-time. The longer the lead-time, the higher
is the average inventory. Lead time has two components, namely the
Administrative or internal lead time (ILT) i.e., from initiation of procurement
action to placing of the order, and the delivery or external lead time (ELT)
from placing the order to deleviry of the material.
3. Order Cycle : The time period between placement of two successive orders is
referred to as a reorder cycle. The reorder may be placed on the basis of two
types of inventory review systems, P-system & Q-system.
4. Stock Replenishment: Replenishment of the stock may occur instantaneously
(when purchased from outside) or uniformly (when the product is
manufactured in-house).
5. Time Horizon : The time period over which the inventory level will be
controlled is the horizon. This may be finite or infinite depending upon the
nature of the demand of the commodity.
6. Re-order Level : The level between maximum and minimum stock, at which
purchasing or manufacturing activities must start for replenishment is called
the reorder level.
7. Re-order Quantity : This is the quantity of replenishment order. In certain
cases it is known as Economic Order Quantity (EOQ).

7.5 Inventory: Significance (Advantages of Inventory)

Inventory is an essential requirement of every facet of business life. There is
hardly any business activity that can be performed without inventory whether it is a
manufacturing concern of good or trading of goods or service organisation like
hospital, banks etc. Even in organisation that processes informations also, the data is
supposed as the inventory and is stored for future references. In whatever form the
inventory is stored, usually, we find the stock points mainly at three stages. These are
at (i) input (raw material) stage (ii) conversion (work-in-process) stage and (iii) output
(finished goods) stage.
In view of the uncertain situations in market, it is physically impossible and
economically impractical to arrive a decision on where, when and how much quantity
of each item is needed. Even if it is possible in a few occassions, it is prohibitively
expensive. This is the fundamental reason in stocking and maintaining the inventory
levels of each item. Thus inventories play a significant role in any organisation. Some
more points to add in the list showing how inventory is advantageous and significant
in an organisation are given below :

1. Inventory ensures continuous and smooth flow of production by providing right

amount of stocks.
366 Operations Research : Theory and Practice

2. It avoids physical impossibility and economical impracticability of getting right

amounts of stock at exact time of need.
3. Inventories enable the organisations to purcahse their materials at
advantageous prices or economic times.
4. Inventory can become a reason for reasonable customer service through supply
without delay. This obviously enhances customer satisfaction and thence the
reputation of the organisation.
5. Carrying inventory makes it possible to maintain more stable operating or
working level.
6. Selection of inventory quantity levels can provide the economy in shipping and
transportation.
7. To plan overall operation strategy through decoupling of successive stages in
the chain of manufacturing or transportation, inventories play a vital role to
minimise the shortage cost and other related costs.
8. To facilitate economic production runs.
9. To facilitate intermittent production of several products on the same facility.
10. Inventories help in achieving favourable Return On Investment (ROI).
11. The decisions regarding effective use of capital is dependent on inventory levels.
12. The efficient usage of available storage space is based on inventory decisions.
13. Make or buy decisions can be effective and economic by proper inventory
policies.
14. Inventory provides means against hedging against future price and delivery
uncertainties.

7.6 Objectives of Inventory

The main objective of inventory control is to minimise the overall investment
(or costs) or inventory carrying at lowest possible level and consistency in operating
requirements. However, the following may also be considered as sub-objectives of
inventory control.
• To minimise carrying cost of inventory.
• To supply the finished product/raw material/ sub-assemblies/semi-finished
goods etc., to the users as per their requirements at right time and at right price.
• To minimise the inactive, surplus, waste, scrap, obsolete, spoilage materials.
• To reduce ttie shortage costs.
• To minimise replacement costs.
• To maximise production efficiency and distribution effectiveness.
• To maximise the reutrn on investment (ROI).
Inventory Management 367

7.7 Functions oflnvenotry

The basic function of inventory is to increase profitability through minimising
the costs of procurement, storage etc.
Since zero-inventory is impractical, the stocking of materials is inevitable and
has a specific objective to fulfill an intended function.
The other functions of inventory are studied under major heads as investment
alternative, geographical specialisation, decoupling, balance between supply and
demand and providing safety stock.

1. Investment Alternative : A major area of asset deployment in an industry is

inventory, since inventory can provide a minimum Return on Investment
(ROI). According to the concept of Marginal Efficiency of Capital (MEC), a firm
should invest in those alternatives, which can provide a greater ROI than capital
cost borrowed. Inventory is one of such alternatives to choose.

In simple words, an investment on the materials can usually be realised to the

extent of investments. And if there is a favourable market or value addition to
the materials, it can be sold at higher price and thus can raise the profitability.
Thus inventory functions as safe guarding investment and also can raise the
ROI.
2. Geographical Specialsation or Territorial Importance : Here, the function of
inventory is to permit geographical specialisation of individual operating units.
Owing to various factors influencing the production and producivity such as
power, water, availability of raw materials / man power, facilities, markets,
means of transportation etc., it is often found economical to locate production
centres considerably away from warehouses. Thus it may become necessary to
collect all the products from each production centre at one place and distribute
to market centres according to demand and shipping costs. Thus inventory
functions to act as mediator between the production and market, on the basis
of geographical specialisation.
3. Decoupling : This function of inventory is to enhance the efficiency of
operations with in a single facility. Suppose, machines^, B and C (say milling,
drilling and grinding) have to run in the order A -» B -> C in a mass production
shop. On any day if machine B is under breakdown, this automatically stops
the machine C, though it is in very good working condition. To avoid this, an
inventory can be maintained in a substore (containing the material already
processed by machine B) between^4 and C, that can ensure smooth production
flow by detaching B and coupling with the substore.

Sub
store
F I G U R E . 7 . 1 : DECOUPLING INVENTORY
368 Operations Research : Theory and Practice

Thus inventory acts as a machine by decoupling in the production chain. This

decoupling also provides the organisaion units to plan and schedule their
activities independently.

4. Balance Between Supply and Demand : Every one is aware of the stocking of
inventory whose supply is seasonal while demand is continuous through out the
year, (e.g., rice, wheat etc)., or demand is seasonal but supply is continuous
(e.g., crackers, rain coats etc). In all such cases, the inventory only can function
as to bring a balance between supply and demand.

5. Safety Stock : Due to short range variation in either demand or replacement

etc., a safety stock or buffer is kept to function a smooth and uniform flow of
production. Safety stock provides protection to the organisation in two types of
uncertainties. First is when the sales (or consumption) is excess of forecast and
second is due to delay in replenishment. Thus inventory acts as a protective
shield from uncertainty of demand.

7.8 Factors Affecting Inventory

1. Economic parameters such as purchase price, production cost, selling price,
procurement cost, carrying cost, shortage cost, operating and information
processing cost etc.
2- Demand.
3. Ordering cycle and its review. (Continuous / periodic)
4. Delivery lead / lag time.
5. Time horizon.
6. Number of supply echelons.
7. Number of stages of inventory.
8. Number of items.
9. Availability and conditions.
10. Government's/Company's policy.

7.9 Concept ofEconomic Order Quantity (EOQ)

It is felt that materials are dead investments unless they are realised into money.
But till then, they block the money and eat away the profits. Therefore materials
managers will be asked to maintain just sufficient stocks. We know well that the higher
stocks block the money while under stocks will interrupt production and makes the
machines/men idle and hence affects the reputation of the industry for not meeting
the demand or not fulfilling the promises. TTius there is a need to arrive a quantity to
be stocked that should neither be high nor low.
Thus the operating doctrine of inventory (i.e., when and how much to order or
stock) depends on how much it costs to the organisation. Thus concept of optimum
order quantity (OOQ) or Economic Order Quantity (EOQ) has gained the prominent
and significant role in inventory management.
Inventory Management 369
The concept of EOQ was first proposed by Ford Wilson Harris in 1913. He has
attempted to bring a trade off between the order costs and holding and shortage costs.
Suppose you are buying some pens for which you have to spend Rs. 100/- to order.
This will be more or less same whether you order 10 or 100. But when you buy 10 the
unit cost is rupees ten and when you buy 100 it is one rupee each. Thus the unit orde
cost decreases on increase of quantity (but never becomes zero).
In case of carrying cost it varies linearly and the cost on storing increases with
increase of quantity. Graphically, it can be observed that the ordering cost er unit
decreases hyperbolically with the increase quantity while the unit carrying cost
increases linearly with the increase of quantity. Thus a trade off between these two can
be obtained to find a most Economic Order Quantity (EOQ) often denoted by Q*
shows most minimum on total cost graph and the normal drawn at EOQ passes
through the point of intersection of order cost curve and carrying cost curve indicating
annual ordering cost is equal to annual carrying cost (i.e., AOC = ACC) at EOQ.

Total
min. cost

EOQ
Costs of simple EOQjtnodel
FIGURE 7 J :

Limitations of EOQ :
1. Ordering to the nearest quantities or packing. Say, instead of ordering 11
dozens, the order may be one gross. This may not be economic.

2. Modifying an order to get a better freight rate. The saving in freight may be
more than compensation the extra holding cost.
3. Simplification of routine becomes difficult: Instead of 13 times a year, one may
order once a month (once in every four weeks gives 13, monthly once gives 12
order)
4. In case of perishable, or bulky items with diminishing consumption or for items
whose market, prices are likely to decline, it may be better to order less than
theoritical order quantity.
5. Seasonal supply factors, market conditions, availability of transport etc., may
indicate larger or smaller purchase quantities. In these situations, judgement
should be given more weight.
6. Liberal discounts or concessional freight rates may suggest larger quantities.
The pros and cons of such purchase should be weighted carefully before taking
the decision.
370 Operations Research : Theory and Practice

7.10 Inventory Models

There are different classifications of inventory models to obtain EOQ. The
models differ based on the variations in nature of demand, allowing or not the
shortages, production or replenishment instantaneously or continuously etc.
However, the models considered in this book are classified as follows :

r
Elementary inventory Quantity discount (price break) Models with restrictions
models models and others

i Probabilistic (stochastic demand) models

Deterministic (known demand) [Note : probabilistic models will be further
models

r
classified and discussed in chapter 3 :
I inventory/production management - II]
Static demand models Dynamic demand models

Estimated maximum lead time and (i) Fixed order quantity system with
normal lead time rate is not uniform variable lead time
(ii) Periodic review system

I I
(No shortage) (Multi - item)
(a) Demand rate D uniform, (With shortages)
(a) Limitation on investment
production rate infinite (a) Scheduling period constant,
(b) Limitation on inventory
(b) Demand rate no uniform but production rate infinite
(stock)
total demand D is prescribed over (b) Scheduling period variable,
(c) Limitation on warehouse
period T, production rate infinite production rate infinite
floor space
(c) Demand rate D uniform, (c) Demand rate D uniform
production rate R is finite production rate R is finite

FIGURE 7 3 :

7.11 Model -1: Wilson Harris Simple EOQ

Assumption In Deterministic Model :
• Demand is known exactly for a given period, usually one year
• Demand is uniform and constant over a period of time.
• No limitations are imposed by storing and clerical capacity.
• The order cost and storage cost is independent of order size.
• Orders are received instantaneously.
• Order and delivery quantities are equal.
• Buffer stock of finished product is independent of order size.
• Price of raw material is stable.
• Order costs per order are same irrespective of order size.
Inventory Management 371
• Set up cost is constant.
• Rate of production is known.
Derivation of EOQ for Constant Demand :
Let Co = ordering cost per order
Cc = carrying cost per unit per year
D = annual demand or requirement per year
p = price of item
Q = ordered quantity per order.
Now,
1. Annual Order Costs (AOC) :
_ . ..- . . Total demand in the year D
Total no. ofr orders in the year = — : -—,,— = —
Quantity ordered per order Q,
.'. Total annual orders cost (AOC)= order cost per order x No. of orders per year

2. Annual Carrying Costs (ACC) :

. . Max. Quantity-min.
l 3quantity 0—0 Q
Average quantity stocked = — •— = -^-r— = ~
Annual carrying costs = carrying cost per unit per period
x average quantity stocked = Q x ^

Total
min. cost

EOQ
i FIGURE 7.4 : ORDER COSTS AND CARRING COSTS AT EOQ

3. Annual Purchase Cost (APC) :

Price item x demand = p.D.
Annual Total Cost (ATC) = (AOC + ACC) + APC
[AOC + ACC are total variable costs while APC is total fixed costs].

ATC =-2-— + Cr.^; + pD

372 Operations Research : Theory and Practice

„ dATC „
For, Q to be optimum, ( Q^ ), ——— = 0

cc
Also TVC = V2C 0 QZ)

Aliter 1: If students do not know differentiation, they may follow the following method
instead of differentiation.
After getting the expressions for/IOC &:^4CC.
As we know that A OC = ACC at EOQ,

We have Co . -^- = C c . — at g ^ ^

Aliter 2 : We can also get the expression by goemetric method.

AOC = No. of orders.xunit cost.
Annual demand
: : x unit order cost
Quantityper order

ACC = sum of areas of all triangles (i.e., inventory cycles) x carrying cost
of each cycle

£ . £ + - g . 2 + . . . + n times I x Cc

= -Q[t + t+ . . . . + ncycles] . C
'c t
2
year]. CC = ±Q.CC

Andat£OQ i.e.,
AOC =ACC

Cc
note: Carrying cost will be usually given in terms ofpercentage ofn interest on unit
price of the item. In such case Cc = i = interest on average inventory andp = price
ofthe item. Similarly, ifthe holding cost (or rent or maintenance cost etc) sayH, is
given along with interest (i) andprice (p) then Cc =H+ip is considered.
Inventory Management 373

iLLUST8ATi0f\i
A Company uses 10000 units per year of an item. The purchase price is Rs. lper
item. Ordering cost is Rs. 25 per order. Carrying cost per year is 12% of the
inventory value. Find
(i) TheEOQ
(ii) The number of orders peryear.
(Hi) If the lead-time is 4 weeks and assuming 50 working weeks per year,
find the reorderpoint
[JNTU(Mech.)98/5]

Solution
Annual Demand (£>) = 10,000 units
Order cost per order (Co) = Rs. 25
Interest on Carrying cost per year (i) = 12 %
Cost of item (p) = Re. 1/-
5 Cycles

2041 Max. qty.

Avg. qty.

2040
Costs of simple EOQ model

FIGURE 7.5 :

10000
(i)

= V416666.6667= 2041.24 = 2000 units.

(ii) The number of orders = —• = ' = 4.899

1 Q 0 Q Q
/ v K A R A
(or) 2QQ0 = 5 orders = 5 orders
No, of working weeks/yr 50
(in) Re order rperiod = — ?—; — = —— = 10 weeks
No. of orders/yr 5
Given Lead time = 4 weeks

/. Re order point is at the end of 6^ week of every cycle.

374 Operations Research : Theory and Practice

ILLUSTRATION 2
A certain item costs Rs. 235per ton. The monthly requirement is 5 tons and each
time the stock is replenished, there is a set-up cost of Rs. 1,000. The costs of
carrying inventory has been estimated at 25% of the value of the stock per yean
What is the optimum order quantity. [JNTU (Mech.) 99]

Solution :
Price (p) = Rs. 235 per ton
13000
Annual Demand (D) = 5 x 12 = 60 tonnes
Order Cost/order (Co) = 1,000 per order
6500
Carrying cost ( Cc) = — • x 60 x 235 (approx.)
1UU
= 1410 9.2

OOQ = FIGURE 7.6 :

2 x 1000 x 60
= 9.2253 tonnes.
1410
ILLUSTRATION 3
A company uses 24000 units of a raw material which costs Rs. 12.50 per unit.
Placing each order costs Rs. 22.50 and the carrying cost is 5.4% per year of the
average inventory. Find the economic order quantity and the total inventory cost
(Including the cost of the material). [JNTU (Mech.) 99/cc (ECE) 99/CCC]

Given data :
Annual demand (D) = 24,000 units 853
cost per unit (p) = Rs. 12.5 per unit
order cost ( Co) = Rs. 22.5 per order
Inventory carrying interest
(i) = 5.4 %. per year of avg. inv.
5.4
= 0.054
100
Inventory carrying cost (Cc)=i.p = 12.5 x 0.054
Economic Order Quantity
2 x 2.25 x 24000
*XC°XD
0.054 x 12.50
= V1600000 = 1264.911 = 1265 units
T o t a l v a r i a b l e c o s t = V2 .CO.CC.D
Inventory Management 375

(i.e carrying costs+ordercosts)= V2 x 22.5 x 0.054 x 12.50x 24000

- V729000 = 853.82
Total cost including material cost
= Total variable cost + fixed cost (i.e., material cost)
= 853.82 + 12.5x24000
= 853.82 + 300000
= Rs. 3000853.82 Ps.
ILLUSTRATION 4

The production departmentfor a company requires 3600 kg of raw materials for

manufacturing a particular item per year. It has been estimated that the cost of
placing an order is Rs. 36 and the cost of carrying inventory is 23 % ofinvestment
on the inventories. The price is Rs. 10 per kg. The purchase manager wishes to
determine the operating doctrinefor raw materials.
So!IAlion :
From the above data, we have (with usual notatios) CQ = RS. 36 per order D =
3600 kg/yr , Cc = 25% of investment i.e. Cc = Rs 10 x 0.25 = 2.5 per kg/yr

5 Cycles

804 Max. qty.

402 -; Avg. qty.

322

32 days = 0.89 yrs

(a) (b)
S FIGURE 7.S :

•-$¥:-
2 x 3600 x 36
(i) The optimal lot size \ - 322 kg/order
2.5
322
(ii) The optimal order cycle time f = y- = ^ =0.89 yr

Assuming 360 days in a year, t* = 32 days

(iii) The total min. variable cost TVC = V2 Co Cc D = V2 x 3600 x 36 x 2.5
= Rs. 804.98 per year.
(iv) Total annual inventory cost = TVC + TFC
= 804.98 + (3600 kg) (Rs. 10%) =Rs. 36804.98 per yr.
-Rs-36805 per yr.
376 Operations Research : Theory and Practice

7*12 Model - 2 .'Continuous and Non-uniform Demand

(Consumption)

Q
'2

Time ^
I FIGURE 7 . 9 ; CONTINUOUS AND NON-UNIFORM CONSUMPTION MODEL

In the above model with usual notations,

Cn.D
We can obtain AOC = - (Same as in model 1)

and ACC = Cc\ |

And at £ 0 ^ i.e., if Q =
Then AOC =ACC
C0D Cc.Q,

c.
[ Thus model-1 and mode-2 are one and same].

Practice Problems
1. An aircraft company uses rivets at an approximate consumption rate of 2,500
kg per year. The rivets cost Rs. 30 per kg and the company personnel estimates
that it costs Rs. 130 to place an order and the inventory carrying cost is 10%
per year. How frequently should orders for rivets be placed and what quantities
should be ordered? [Maduri B.E. (Mech) 1980, JNTU (MeCh.) 94/CCC]
Answer: Q* = 4^ ^ r i v e t s ;T* = 0.18 years
2. A manufacturer has to supply his customer with 600 units of his product per
year. Shortages are not allowed and the storage cost amounts to Rs. 0.60 per
unit per year. The set-up cost per run is Rs. 80. Find the optimum run size and
the minimum average yearly cost. [Marathwada, M.Sc. (Math), 1982, JNTU (CSE) 95/S
Answer: Q* = 400 un[ts -t* = 8 months TVC = Rs. 240
Inventory Management 377

3. A company works 50 weeks in a year. For a certain part, included in the assembly
of several parts, there is an annual demand of 10,000 units.
This part may be obtained from either an outside supplier or a subsidiary
company. The following data relating to the part are given :
Description From Outside Supplier From Subsidiary Company
Rs. Rs.
Purchase price/unit 12 13
Cost of placing an order 10 10
Cost of receiving an order 20 15
Storage and all carrying cost,
including capital cost per unit 2 2
per annum
(i) What purchase quantity and from which source would you recommend?
(ii) What would be the minimum total cost [ICWA, June 1989]
Answer : (i) Q* = 547,72 units TVC* = Rs. 1,21,095.45
(ii) <2* = 500 units ; TVC* = Rs. 1,31,000 Purchase must be made from an
outside supplier.
4. A manufacturing company uses certain part at a constant rate of 4,000 units
per year. Each unit costs Rs. 21- and the company personnel estimates that it
costs Rs. 50 to place an order, the carrying costs of inventory is estimated to be
20% per year, find the optimum size of each order and minimum yearly costs.
[Bharathidasan M.Sc. (Math) 1981, 0U (Mech.) 88]
Answer: Q* = 1000 units, TVC = Rs. 400
5. Robert has to supply his customers 100 units of the product 'X' only every Friday
and only then he obtains 'X' from Garry, his local supplier, at Rs. 60 per unit.
The cost of ordering and transportation from Garry is Rs. 150 per order. The
cost of carrying inventory is estimated at 15% per year of the cost of the product
carried.
(i) Describe graphically the inventory system
(ii) Find the lot size which will minimize the cost of the system,
(iii) Determine the optimal cost. [JNTU (Mech.) 93/MS]

Answer: (n) (£ = 416units; (iii) TC* = Rs. 6072

6. Sainath keeps his inventory in special containers. Each container occupies 10
Sq. ft. of store space. Only 5000 sq. ft of the 9000 containers, priced Rs. 8 per
container. T h e ordering cost is estimated at Rs. 40 per order, and the annual
carrying costs amount to 25% of the inventory value.
Would you recommend Sainath to increase his storage space? If so, how much
should be increased? [Delhi Univ., M.Com. 1990, 3NTU (EEE & ECE) 95/MS]
Answer : Increase in storage space = 1000 sq. ft.
378 Operations Research : Theory and Practice

7. Find the economic lot size, the associated total cost and the length of time
between two orders, given that the set-up cost is Rs. 100, daily holding cost per
unit of inventory is 5 paise and daily demand is approximately 30 units.
[JNTU (CSE) 92]
Answer : Q* = 3 4 6 units> jyC* = 173; / = 11.5 days
8. Shivnath company buys in lots of 2000 units which is only 3 months supply.
T h e cost per unit is Rs. 125 and the order cost is Rs. 250. T h e inventory carrying
cost is Rs. 20% of unit value. How much money can be saved by using economic
order quantity? [JNTU (Mech. & ECE) 95/CCC]
Answer: Rs. 16000/-
9. A ship-building firm uses rivets at a constant rate of 20,000 numbers per year.
Ordering costs are Rs. 30 per year. Each rivet costs Rs. 1.50 and the holding
cost is estimated to be 12.5% of unit cost p e r unit per year. Determine EOQ.
[JNTU (Mech.) 93/MS]
Answer; Q* = 1005 rivets
10. A computer manufacturing company wishes to know what E O Q should be
ordered for its chips. T h e average daily requirement is 120 units and the
company has 250 working days a year- T h e manufacturing cost is 50 paise per
part. T h e sum of annual rate of interest, insurance, taxes and so forth is 20%
of the unit cost and the cost of preparation is Rs. 50 per lo£JNTU (CSE) 94/S/MS]
Answer: g* = 5477 units
11. Consider the inventory system with the following data in usual notations:
D = 1000 units/year ; i = 0.30; p = Rs. 0.50/unit; C o = Rs. 10.00; L = 2 years
(lead time).
(i) Optimal order quantity
(ii) Reorder point;
(iii) Minimum average cost. [JNTU - CSE/ECE 2001]

713 Model 3: With Continuous Production (Supply) Rate

This model is similar to that of model-1 except that of instantaneous
replenishment. Instead of instaneous replenishment, this model assumes to replenish
continuously till the maximum quantity is reached due to the fact that in many
situations, the amounts ordered will not be received all at once but will be available at
finite supply rate. This order may be produced internally also.
Thus in addition to the usual assumption the following points are assumed in
this model.
1. Supply is continuous till Qmax *s reached and then it stops.
2. The production rate or supply rate (say r) is greater than demand or
consumption rate (d) i.e., r > d.
3. Production begins immediately after production set-up.
4. During production also consumption will be there and when Qmax r e a c n e s >
production stops while consumption continues till Qmin *s reached.
Inventory Management 379
Graphically this can be shown as follows :

Amount produced

I Amount consumed, during production

Imax tp(pd) Q
vg n
_ "7^^" -

| Inventory depletion \*—

\* Production cycle time •]

FIGURE 7 . 1 0 : CONTINUOUS SUPPLY RATE MODEL

In the above model,'dtp= production time and r = rate of production.

We have the quantity produced

Q, =r.tp => tp = &

During the production time, the inventory increases at the rate of r, but due to
consumption simultaneouly this inventory decreases at the rate old.

Tlierefore the inventory accumulates at the rate of (r - d).

Thus inventory accumulation
/ m a x = rate of accumulation x production time.

and we have / m i n = 0.
[Since we consume till all the inventory is exhausted before starting new
production cycle].
Thus average inventory = - (7 max + / m i n j = & i-d-
r
Now, annual carrying cost (ACC) - Cc . average quantity

And we know that

Annual ordering costs (AOC) = Co . •—

380 Operations Research : Theory and Practice

Total variables costs (JVC) = Co — + Cc. ^ I 1 - -

Or otherwise, at EOQ AOC=ACC

d

Other Important Formulae :

1. Total variable cost

2
6oPt ° 7"

(r-d
or
2. Optimal length of production run

3. Optimal production cycle time

D D y Cc \r~d
4. Optimal number of production runs

^cD(r-d

Remarks :
In all the above formulae, there may be a confusion between D and d. D is
annual demand or annual consumption while d is consumption rate. If there is only

one cycle per year, the formula can be written as g ^ = A/ ——— —.

Note : It is more appropriate to use the term EBQ(Economic Batch Quantity) instead ofEOQ (Economic
Order Quantity) for Qppi and 'set-up cost' is more appropriate term in the place of 'order cost3
(Co). Therefore these words are often found in the problems of this model.
Inventory Management 381

ILLUSTRATION 5
A contractor has to supply 10,000 bearings per day to an automobile
manufacturer. He finds that when he starts production run, he can produce
25,000 bearings per day. The cost of holding a bearing in stock for a year is Rs. 2
and the setup cost of a production run is Rs. 180. How frequently should
production run be made? Also find production run time and total variable cost
(Assume 300 days in the year). [JNTU (Mech.) 97/p]
Solution :
According to the problem, with usual notations,
Set-up cost (Co) = Rs. 180 per production run.
Carrying cost (Cc) = Rs. 2 per bearing per year.
Rate of production (r) — 25, 000 bearings per day.
Rate of consumption (d) = 10,000 bearings per day.
Annual demand = 10,000 x 300 = 30,00,000 / year
(Assume 300 days in a year)

K= 1000 units 180 K

d = days

90 K -

Q0 = 30K K=1000
Costs of simple EOQ model
(b)
(a)
FIGURE 7.-II :
Now, knowing all the terms, we can find EBQ (Economic Batch Quantity)

-d

-V 2 x 30,00,000 x 180
2
25000
(25000 - 10000)

= -y 54 x 107 x y^ = V90x 107 = 30000 bearings.

Frequency of production run is given by-t*.
Qjopt 30000

Production cycle time

• Qopt 30000 6 1
l 1
382 Operations Research : Theory and Practice

Total variable cost = Co . — + Cc. — . \ —

or V 2CODCC
•-d

25000-10000
TVC
•V 2x 180 x 3000000 x-2 x 10000

•^-««""S
= \ / 4 x 5 4 x 108x 11
10
= Rs. 1, 80,000
ILLUSTRATION 6
The manager of a company manufacturing car parts has entered into a contract
of supplying 1000 numbers per day of a particularpart to a car manufacturer. He
finds that his plant has a capacity of producing 2000 numbers per day of the part.
The cost of the part is Rs.50. Cost of holding stock is 12% per annum and set up
cost per production run is Rs. 100. What should be run size for each production
run and total optimum cost/month? How frequently should production runs be
made? Shortage is notpermissible. [JNTU (CSE) 97/S]

Solution
Demand rate = Quantity to be supplied per day (d)= 1000
Production rate = Quantity produced per day (r) = 2000
cost of the part(p) = Rs. 50
Interest on Holding cost (i) =12% of annual cost.
Holding cost (Cc=ip) = 0.12 x 50
Set up cost (Co) = Rs. 100 per production run.
No. of working days is assumed as 365 days.
.-. Demand (D) = 365x1000

5000

5000
5 days Costs of simple EOQ model
(a) (b)
FIGURE 7 . 1 2 :
Inventory Management 383

Optimal Run Size

/2x365000x 100 ( 2000 ^

= V 0.12x50 I 2000-1000 j
= 4932.9 = 5000 parts.

Frequency of production run (t*) = ^*- = = 5 days.

Total variable cost = *y 2 . Co. C c . D 1 - -

!x 100x0.12x50x356000 1-

= Rs. 14798.65 per year

= Rs. 1233.22 per month

365Q( x 5Q
Total fixed cost/month = g = 15,20,833.33

Total optimum cost per month — Rs. 1522066.55

Practice Problems
1. (a) A product 'X* is purchased by a company from outside suppliers. T h e
consumption is 10,000 units per year. T h e cost of the item is Rs. 5 per unit
and the ordering cost is estimated to be Rs. 100 per order. T h e cost of
carrying inventory is 25%. If the consumption rate is uniform, determine
the economic purchasing quantity.
(b) In the above problem assume that company is going to manufacture the
item with the equipment that is estimated to produce 100 units per day.
The cost of the unit thus produced is Rs. 3.50 per unit. T h e set-up cost is
Rs. 150 per set-up and the inventory carrying charge is 25%. How has the
answer changed ? [JNTU (Mech. & ECE) 92/S/CCC]

Answer; (a) Q* = 1,265 units (b) Q* = 2,391 units

2. An item is produced at the rate of 50 items per day. T h e demand occurs at the
rate of 25 items per day. If the set-up cost is Rs. 100 per setup and holding cost
is 0.01 unit of item per day, fmd the economic lot size for one run, assuming
that the shortages are not permitted. Also fmd the time of cycle and minimum
total costs for one run. [0U (Mech.) 91/S]

Answer : Q* = j 0 0 0 ltems • j * = 40 days; TVC* = Rs. 5 per day.

Total cost per run = Rs. 200

384 Operations Research : Theory and Practice

3. An item is produced at the rate of 128 units per day. T h e annual demand is
6,400 units. The set-up cost for each production run is Rs. 24 and inventory
carrying costs is Rs. 3 per unit per year. There are 250 working days for production
each year. Develop an inventory policy for this item. [IGNOU (MBA) 94]
Answer: Q* = 353 units, t* = 14 working days, 7^ = 2.8days TC = Rs. 858.65 per yer

4. T h e annual demand for a product is 1000000 units. T h e rate of production is

2000000 units per year. T h e set-up cost per production run is Rs. 5000 and the
variable production cost of each item is Rs. 10. T h e annual holding cost per
unit is 20% of its value.Find the optimum production lot size, and the length
of the production run. [IGNOU (MBA, Assignment) 95]
Answer : Q* = 31600 units ; t* = 115 days
5. You have been given the following information regarding the production lot
size of a particular p r o d u c t :
Annual demand = 5,000 units
Set-up cost = Rs. 100 per set-up
Daily demand = 1 7 units.
Production rate = 50 units per day
Optimum production lot size = 275 units.
Rising interest rates have caused a 10% increase in the holding costs. Determine
the new optimum production lot size for the product.
[IGNOU (MBA, Assignment) 96]

6. Assuming you are reviewing the production lot size decision associated with a
production operating where production is 8000 units a year, annual demand
is 2000 units, set-up costs is Rs. 300 per production r u n and holding cost is Rs.
1.60 per unit per year. T h e current production run is 500 units every 3 months.
Would you recommend a change in the production lot size ? If so, why? How
much could be saved by adopting the new production run lot size?
[JNTU (Mech.) 90/S/P]
Answer: 750
7. A contractor has to supply 20,000 units per day. He can produce 30,000 units
per day, the cost of holding a unit stock is Rs. 3 per year and the set-up cost
per run is Rs. 50. How frequently, and of what size, the production runs be
made?
[JNTU (Mech.) 92/S/P]
Answer : Q* = 1414 u n i t s ; f = L 6 8 hrs
8. Find the most economic batch quantity of a product on a machine if the
production rate of the item on the machine is 200 piece/day and the demand
is uniform at the rate of 100 pieces/day. The set-up cost is Rs. 200 per batch
and the cost of holding one item in inventory is Rs. 0.81 per day. How will be
batch quantity vary if the machine production rate was infinite?
[JNTU (CSE) 94/S]

Answer: 157
Inventory Management 385

7.14 Model - 4: Deterministic Model With Shortages

So far we have seen the models with an assumption that no shortages are
allowed, and the EOQ derivation involved a trade-off between ordering cost and
carrying cost. But, there could be situations where shortages may occur or may be
allowed to gain some advantages.
Advantages of Allowing Shortages :
The following advantages can be derived by allowing shortages.
1. By allowing shortages, the cycle time increases, and hence the order (or setup)
cost can be spread over longer period.
2. If unit value of the inventory is high, carrying costs also will increase. In such
cases the shortage cost could be less than the increased or differed carrying
cost.

[As ^ AB x CE > ^ ED x CE
AB > ED]

^"-shortage D

FIGURE 7 . 1 3 : MODEL WITH SHORTAGES

In the above graph, if shortage is allowed, the graph takes the form ABCDE
and if not allowed it takes the form ABE. Now, we can observe that area of BCE (l/2
AB x CE) i.e., difference in carrying cost is greater than the area covered by
A CDE (V2 ED x CE), i.e., shortage cost.

[This is true as long as the shortage quantity is safe i.e., ED <AB]

Thus to encash the above advantages, managers often choose to have shortages;
However, the shortage should not be permitted if it is expected to cause heavy loss of
production or damage.
Now, let us derive the expression for EOQ for this model.
Expression for EOQ :

This model is based on all the assumptions of Model - 1 except that inventory
can be allowed to run out of stock for a certain period of time. Usually in business
situations the following two types of such running out of stock can occur.
1. Sales may be lost due to customers are not likely to buy the inventory items.
2. Customers may leave orders with the supplier and this back order is filled on
stock availability.
386 Operations Research : Theory and Practice

Thus all the costs such as cost of keeping backlog reorders, loss of good will,
cost of shipping to customers etc., depend on how long a customer can wait to receive
the order.
In deriving the expression for EOQ of this model, the following terms are used.
^ = time between receipt of the order and when the inventory level
reaches zero.
ts = Shortage time or stock - out time or back order time.
t = total cycle time = t- + ts
Q = Number of units that are back ordered or shortage quantity.
Q = Max. inventory quantity (when received)
Q = Total quantity = Q. + Q
Cs ~ Shortage costs, and other terms are as usual.
Co = Order cost per order
Cc = Carrying cost
i = Interest on average carrying inventory
p = Price of the item
D = Annual demand

FIGURE 7. i 4 : MODEL WITH SHORTAGES

In this model, except for the fixed costs i.e., procurement cost (D.p), all the
costs will be affected. Therefore, we need to determine the optimal values of order
quantity Q ^ . (or £>*), optimal stock level (Q$) and optimal number of shortage units
that can be allowed (Q^). And hence the objective function for this model would be,
to minimise the total variable cost (TVC), where
TVC = Annual oder cost + Annual carrying cost + Annual shortage cost
= AOC + ACC + ASC
Now, we know that,
Quantity of units consumed
Demand (D) =
Consumption period

t D
Inventory Management 387

c- -, , & Qs Z-Qi , Q,
Similarly, h = ~f) > ^ = TT = —r>— anc
* ^ = "n
Also, from the above graph
A ABC and A CDE are similar
AC__AB_ ^ h_ _ _ ^
C£ " DE "' Qi ~ ! i

also —=—

The average inentoiy during the period ^ , is the area of A ABC = — - Q • t{

l Qi-t <8
Average carrying inventory = - ft • - — = — -t

Carrying cost =—-t-Cc

4
per cycle ^
o

Annual carrying cost - —— • Cc

(t - 1 year) ^

Similarly, for the period ts, the average inventory is given by the area of A CDE

:. Average shortage inventory = —• • CEx ED

D
• • ' t

Shortage cost per cycle = — — t • C5

Annaual shortage cost = ———

(ft-ft) 2 Cs
(t = 1 year)
Cn-D
And annual order costs = -

2Q.
388 Operations Research : Theory and Practice

Thus the above relation states that TVC is the function of two variables
Q and Q^. According to the principles of partial derivative, the minimum values of
Q and Q may be calculated with the conditions that

a2 (TVC) a2 (TVC) _ \d2 (TVQ)

On solving the above equations we get, optimum order quantity.

ZC D c s
EOQ =0 =-\l ° A \

* / 9 z"1 n / z"1 A
and optimal stock level

f C i / 2CQZ) C
and optimal shortage level QT =• Q* ~ Ql = Q* I ^ I = A/ ^

Cs
also TVC* = - \ / o ^ n n (

2 C
and total cycle time (t) = ^- = <\ / Q (Cc + i
D
V D C C ' [ CS
Total cost TVC + TFC = TVC + pD
ILLUSTRATION 7
A manufacturer has to supply his customer 30,000 units ofproduct/year. Demand
is known andfixed. There is no storage space and shipping is daily. The penalty
for failure to supply is Rs. 0.20 per unit per month. Inventory holding cost is
Rs. 0.1 per month and set up cost is RsJSOAper production run. Find optimum lot
sizefor the manufacturer. [JNTU (CSE) 98]

Solution :
D = 30000 units/yr ('D' is fixed hence deterministic).
There is a penalty for failure to supply.
Therefore shortage cost (Cs) = Rs. 0.20 per unit per month.
= 0.2 xl2 = Rs. 2.40 per unit per year.
Holding cost = Rs. 0.1 per unit per month
Holding or carrying cost (Cc)= 0.1 xl2 = 1.2 per unit per year. Order cost or
set-up cost (Co) = Rs. 350 per prodn. run.
Inventory Management 389

36000

/2x350>
350x30000 3.6
X 18000
2 2.4

/3500x
5000 or 5500 (or 5250)
Costs of simple EOQ model
= 500 x 10 - 5000 or 500 x 11 = 5500 FIGURE 7 . 1 5 :
ILLUSTRATION 8 „___________ _____ —-
A commodity is to be supplied at a constant rate of 200 units per day. Supplies of
any amount can be had at any required time, but each ordering costs Rs. 50; costs
of holding the commodity in inventory is Rs.2.00per unit per day while the delay
in the supply of the items induces a penalty ofRsAOper unit per day. Find the
optimal policy (Q, t), where ftf is the reorder cycle period and Q is the inventory
level after re-order. Also find the optimal inventory level and shortage units. What
would be the bestpolicy, if the penalty cost becomes infinity? [JNTU (CSE) 95/S]

Solution :
From the data of problem, with usual notations, we have
D = 200 units/day Co = Rs. 50 per order
Cc — Rs. 2 per unit per day Cs = Rs. 10 per unit per day

If shortages are allowed

Q5 = 91.3

/RE-order

o
109.5
Q s = 18.2

0.547
Shortages
(a) (b)
FIGURE 7 . 1 6 :
Now,
jy
(i) Optimal order quantity

2 x 50 x 200 f 2 + 10
- 109.5 units
10
390 Operations Research : Theory and Practice

(ii) Optimal inventory level

r2x = 91.3 units

2 | 2 + 1.0
(iii) Optimal shortage units (g = £* - Q*
= 109.5-91.3=18.2
0* 109 5
(iv) Reorder cycle time ** = 77 = 9nn = ° ' 5 4 7 day
200

(v) Total veriable cost

10
=\ = Rs.l82.57
2 x 50 x 2 x 200
2+10
(vi) If penalty cost Cs = «», the shortage cost is unaffordable i.e., shortage should not
be allowed, thus the expression for Q* becomes

Im
If shortages are not allowed
00

0 0.5 days
k
(a)

FIGURE 7 . 1 7 :

2 x 50 x 200
= 100 units

100 1
and
Total variable cost - V2 • Co • Cc • D = V2 x 50 x 2 x 200 = Rs. 200/-

Practice Problems
1. The demand of an item is uniform at a rate of 25 units per month. The Fixed
cost is Rs. 15 each time of production run made. The production cost is Rs. 1
per item, and the inventory carrying cost is Rs. 0.30 per item per month. If the
shortage cost is Rs. 1.50 per item per month determine how often to make a
production run and of what size is should be? [Osmania M.Sc. (Stat), 1984]
Answer : Q* = 54 items ; t* = 2.16 months
Inventory Management 391

2. Rajiv automobiles has to supply its customers 24,000 units of its product per
year. This demand is fixed and known. The customer has no storage space and
so the manufacturer has to transport a day's supply each day. If the
manufacturer fails to supply, the penalty is Rs. 0.20 per unit per month. The
inventory holding cost amounts to Rs. 0.10 per month and the set-up cost is
Rs. 350 per production run. Find the optimum lot size of the manufacturer.
[0U (M.Sc.) 1985]

Answer : £* = 4,578 units per run

3. The cost parameters and other factors for a production inventory system of
automobile pistons are given below. Find : (i) Optimal lot size ; (ii) Number of
shortages ; (iii) Manufacturing time and time between set-ups.
Demand per year = 6000 units
Set-up costs = Rs. 500
Molding costs per year = Rs. 8
Unit cost = Rs. 40
Production rate per year = 36,000 units, shortage cost per unit per yr - Rs. 20.
[IGNOU (MBA) 96]

Answer: (i) £f = 1,123 units (ii) (£ = 266.6 units

(iii) t\ = 0.03 per year; t* = 0.19 year

4. The demand for an item is deterministic and constant over a time and it is equal
to 600 units per year. The per unit cost of the item is Rs. 50 while the cost of
placing an order is Rs. 5. The inventory carrying cost is 20% of the cost of
inventory per annum and the cost of shortage is Re. 1 per unit per month. Find
the optimal ordering quantity when stockouts are permitted. If the stockouts
are not permitted, what would be the loss to the company. [ICWA, Dec. 1985]

Answer: (i) Q* «• 33 units ; R* = 15 units ; TVC = Rs. 181

(ii) £* = 24.5 units; TVC = Rs. 245 loss = Rs 64

5. The demand for a purchased item is 1000 units per month, and shortages are
allowed. If the unit cost is Rs. 1.50 per unit, the cost of making one purchase is
Rs. 600, the holding cost for one unit is Rs. 2 per year, and the cost one shortage
is Rs. 10 per year.
Determine :
(i) The optimum purchase quantity
(ii) The number of orders per year
(iii) The optimal total yearly cost.
(iv) Represent the model graphically. [Bangalore ME 1980]
Answer: (}) Q* = 3 6 00 units, (ii) 50 order/yer (iii) TVC* = Rs. 6000
392 Operations Research : Theory and Practice

7.15 Model 5 :With Finite Replenishment and Shortages

Allowed
This model is a combination model -3 and Model - 4. The graph for this model
will be as follows.

Inventory

t
a
o
t
6
Shortage
- Production »\« Shortage
g
cycle cycle

FIGURE 7 . 1 8 ; MODEL WITH SHORTAGES AND CONTINUOUS SUPPLY RATE

*i = period with production and consumption
£2 = period of consumption
% = period of consumption with shortage
t^ — period of production is shortage and consumption
The expressions for various terms can be derived by clubbing the model - 3 8c
4 and we get the following formulae .

1. Optimal production lot size

r-d c
I °)
2. Optimal inventory level Q* =

3. Optimal level of shortage

-4 2C0D

(£ - Q* •
,r_(

-V 2C 0J D

4. Production cycle time

D D y C \r-d\\ C

r-rfA
Inventory Management 393

5. Total min. variable cost

1. If r = oo, the model becomes same as that of instant replenishment with

shortagei.e., model- 4.
2. IfCs = co, the model becomes same as that of finite replenishment (prodn. rate)
without shortage Le.y model -3.
3. IfCs=oo andr=oo, then the model becomes instaneous replenishment without
shortages ie., model- L
Note : Students are advised to remember the formulae of this model and may deduce to any other models
using the points 6 to 8 given above.
ILLUSTRATION 9 .
An automobile company uses 6000 pistons per yean The company can
manufacture the pistons at the rate of 36000 units per year with a set-up cost of
Rs. 2000. The cost of holding inventory per year is estimated to be Rs. 8/-per unit
cost is Rs. 40. If the company has a provision to allow shortage at the cost ofRs.
20 per unit per year, find (a) Optimal lot size (b) No. of shortages (c)
Manufacturing time and (d) time between set-ups, (e) also find total cost
(including material cost) [JNTU(Mech. 92/s/P]
Solution :
The given data in the problem, with usual notations is
Demand (D) =6000 units/year (consumption rate)
Production rate (r) = 3600 units/year
Set-up costs (Co) = Rs. 2000 per set-up
Carrying cost (Cc) = Rs. 8 per unit per year
Shortage cost (Cs) = Rs, 20 per unit per year
Price (p) = Rs. 40
(Here d = D since it is given in terms of 'per year')
C0D f r )K
(a) Economic lot size (ft*)
C. [r A
x2000 x 6000 |(
cs
36000
}
r
"| I 8 + 20"|
8 ^36000-6000 J | 20
<• J
= 2,246 units
cc V
(b) No. of shortages <£)
3600
8 + 2<
394 Operations Research : Theory and Practice

0* 2246
(c) Manufacturing time t\ = — = = 0.06 year
= 0.06x365 = 21.90 days
~ 22 days
0* 2246
(d) Time between set-ups t* =y-= 9 4 n n f l = 0.38 year
= 0.36x365= 138.7 days

(e) Total cost =— +- —+ — +pD

2x2000,8,6000,1-1^00
1^
9 2 x - x - x 108 +24000
6 7
= 10690.45 + 240000
= 250690.45

Practice Problems
1. The demand for an item in a company is 18,000 units per year, and the
company can produce the item at the rate of 3,000 per month, the cost of one
set-up is Rs. 500 and the holding cost of one unit per month is. 15 rupees. The
shortage cost of one unit is Rs. 240/- per year. Determine the optimum
manufacturing quantity and the number of shortages. Also determine the
manufacturing time and the time between set-ups.
Answer : 4,489 units, 17 units, 1.5 months, 3 months
2. The demand for a product is 25 units per month and the items are withdrawn
uniformly. The set-up cost each time a production run is Rs. 15. The inventory
holding cost is Rs. 0.30 per item per month.
(i) Determine how often to make production rim, if shortages are not
allowed?
(ii) Determine how often to make production rim, if shortages cost Rs. 1.50
per item per month [Madras B.E. 1990]
Answer: (i) Q* = 50 (ii) 54.7 units
Inventory Management 395

3. T h e demand of an item is uniform at a rate of 20 units per month. T h e fixed

cost is Rs. 10 each time a production run is made. T h e production cost is Rs 1
per item and the inventory carrying cost is Rs. 0.25 per item per month. If the
shortage cost is Rs. 1.25 per item per month, determine how often to make a
production run and of what size should it be? [Osmania M.Sc. (Stat) 1984,
Mysore B.E. (Mech) 1985]
Answer : Q* = 44 i t e m s ; t* = 2.2 months
4. A company has a demand of 12000 units/year for an item and it can produce
2000 such items per month. T h e cost of one set-up is Rs. 400 and the holding
cost per unit/month is Re 0.15. T h e shortage cost of one unit is Rs. 20 per year,
find the optimum lot size and the total cost per year, assuming the cost of 1
unit is Rs. 4. Also find the maximum inventory manufacturing time and total
time.
[JNTU (Mech.) 96/MSj
Answer : Q* = 3 4 ^ u n i t s TVC* = R s , 5 1 3 3 6
5. T h e demand for an item in a company is 15000 units per year and the company
can produce the items at a rate of 300 per month. T h e cost of one set-up is Rs.
500 and holding cost of 1 unit per month is 15 paise. T h e shortage cost of one
unit is Rs. 20 per month.
Determine
(i) Optimum production batch quantity and number of shortages
(ii) Optimum cycle time and production time
(iii) Maximum inventory level in the cycle
(iv) Total associated cost per year if the cost of the items is Rs. 20 per unit.
[JNTU (Mech.) 94]
Answer : ( i} Q* = 4)489 g = n units (ii) L 5 ) 3 mont hs
(iii) 2227 units (iv) Rs. 334/-

7.16 Model 6: With Price Breaks or Quantity Discounts

In the earlier models it is assumed that the cost of offered procurement is
unaltered by order size, but, in some situations the discounts will be ofered to purchase
more units to encourage the buyers. In such occassions, we have differential prices for
different quantities ordered. Thus the decision whether the price discounts should be
availed or not is another important aspect of inventory policy. The computation
method is given below.
Model with 2 - Price Breaks :
Suppose the following price discounts are offered where b1,b2 are the limits of
discounts, and ch c^ c% are the costs respectively.
Mathematically,
Quantity Price Per Unit
0<QA<b{ cx
396 Operations Research : Theory and Practice

Notice that c3 < c2 < c\

b, b2
Q—
FfGURE/,19: MODEL WITH PRICE BREAKS

Now, for the above problem, with usual notations, the algorithm is as follows.
The Algorithm :

Step 1: Consider the lowest price i.e., c3 and determine Q3 with usual EOQ formula

i.e., Qs

Now, if Q| > b2, then Qs is the EOQ for the given problem i.e., Q* = Q^ and
hence optimal TVC%, (for Q3) can be computed. If Q3 < b2, then go to step - 2.

/ 1 Co D
Step 2 : Consider c2 a n d compute Q% with the EOQ formula i.e., Q% = A/

Now if Ql n e s in between ^j and b% i.e., if ^j < Q| < ^2 ^s true, then compare

TVC* (at (g) and TVC (at ft2)- If TVC* (at Q|) < TFC (at b2) then g* = (&> i-e.,
Q2 is the EOQ, otherwise Q* = ^ is the EOQ required.

< ^1 then go to step 3.

Step 3 : Calculate Qj based on c1 and compare TVC (for fej), TFC (for b2) and
TFC ( for Qj) to find EOQ. Obviously the quantity with the lowest cost will be
the required EOQ.
Note : Thus model can be extended to any number ofprice - breaks in the same pattern as above steps.

The above problem can also be computed by calculating EOQs for various conditions separately
and comparing the corresponding costs.
Inventory Management 397

ILLUSTRATION 10
Monthly demand for an item is 200 units. Ordering cost is Rs. 350, inventory
carrying charge is 24% of the purchase price peryean The purchase prices arePt
=Rs. 10for purchasing Qt < 500; Pt =Rs. 9.2 5for purchasing 500 < Q2 750 andP3
Rs. 8.75for purchasing 750 < Qs.
Determine optimum purchase quantity. If the order cost is reduced to Rs. 100 per
order, compute the optimum purchase quantity. [JNTU (Mech.) 95, IGNOU (MBA) 96]

S o I uf ion :
Case (i)
Co = 350

D = 200 x 12 = 2400
Cc = 0.24 Pi

Ql = V 77 where i = 1 , 2 , 3 . . . .

: 350 x 2400 onA .

= 894 units.
0.24 x 8.75
Since 894 is greater than 750, the optimum purchase quantity is 894.
Case (ii) :
Cn = 100

12 = 2400
and C c = 0.24/;-where i= 1,2, 3.

x 100x2400
0.24 x 8.75
since 478 < 750 = b%, we next compute

2 x 100 x 2400
0.24x9.25 =
since 465 < 500 = &2, we next compute

V 9 v 1 DO v 9400
0.24x10 ^447unitS>Q-
We, now compare the total costs for purchasing.
SI = 447
b2 = 500 and
Z>3 = 750 units respectively.
398 Operations Research : Theory and Practice

TCX (for purchasing 447) = 10 x 2400 + JJWiSji2° + I ( 0 24) x 10 x 447

2
= Rs. 25,085

10QX
TC2 (for ^ = 5 0 0 ) = 9.25 x 2400 + ^ 4 Q Q + ^ (0.24) x 9.25 x 500

= Rs. 23,247

1QQx
TC$ (for Q3 = 750) = 8.75 x 2400+ 2 4 Q Q + I (0 .24) x 8.75 x 750
750 I
= Rs. 22,119.50*

Therefore economic purchase quantity in this case is Q$ * = 750.

Practice Problems
1. Find the optimum order quantity for a product for which the price breaks are
as follows ;
Quantity Unit Cost (Rs.)
0< ( 2 ^ 8 0 0 Re 1.00
800 < Q2 Re 0.98

The yearly demand for the product is 1,600 units per year, cost of placing an
order is Rs 5. The cost of shortage is 10% per year. [0U (MBA) 88]
Answer: Q* == g00 units

2. The annual demand of a product is 10,000 units. Each unit costs Rs. 100 if
orders are placed in quantities below 200 units but for orders of 200 and above
the price is Rs. 95. The annual inventory holding costs is 10% of the value of
the item and the ordering cost is Rs. 5 per order. Find the economic lot size.
[JNTU Mech/Chem/Prod - 2001/S, 0U - (MBA) 89]
Answer: Q* = 100 units

3. Find the optimum order quantity for the following

Annual demand = 3600 units ordering costs = Rs. 50
Cost of storage = 20% of unit costs
Price Break :
Quantity Unit Cost (Rs.)
0 < (?i < 100 20
100 < Q,2 18
[JNTU (CSE) 93]
Answer: O* = 316.23
Inventory Management 399
4. Find the optimal order quantity for a product for which the price - breaks are
as follows :
Quantity Unit Cost (Rs.)
Below 100 200
101 to 200 180
above 200 160
Monthly demand of the product is 400 units. The storage cost is 20% per year,
of the price of the product per unit. Ordering cost is Rs. 50 per order.
[0U, B.E. (Prod) 1993]
Answer: Q* = 2 00units
5. Find the optimum order quantity for a product for which the price breaks are
as follows.
Quantity Unit Cost (Rs.)
:100 20
< 200 18
200 < 16
The monthly demand for the product is 400 units. The storage cost is 20% of
the unit cost of the product and the cost of ordering is Rs. 25 per month.
[JNTU (ECE & EEE) 94]
Answer: Q* = 2 00 units
6. A manufacturer of engines is required to purchase 4,800 casting per year. The
requirement is assumed to be known as fixed. These castings are subject to
quantity discounts. The price schedule is as follows.
Quantity Unit Cost (Rs.)
Less than 500 150
500 to 750 138.75
750 or more 131.25
Monthly holding cost expressed as a decimal fraction of the value of the unit is
0.02- Set-up costs associated with the procurement of purchased item is Rs. 75
per procurement. Find the optimum purchase quantity per procurement.

Answer: (2* = 1656units

7. Annual demand for an item is 500 units odering costs is Rs. 18 per order.
Inventory carrying cost is Rs. 15 per unit per year relationship between price
and quantity ordered is as follows :
Quantity Ordered : 1 to 15 16 to 149 150 to 549 550 & above
Price Per Unit (Rs) : 10 9 8.75 8.5
Specify optimal order quantity and the corresponding price of this item.
[I.I.I.E. Grad. 1982]

Answer : Q* = i 6 units j y c * = Rs. 6142.50

400 Operations Research : Theory and Practice

8. A diesel engine manufacturing company has planned its production schedule

for the next year based on the forecasted demand, backorders and plant
capacity. Instead of manufacturing the piston, that goes into the final product,
the company has decided to buy the pistons. The pistons are required at a rate
of 60 per day. Ordering costs have been estimated at Rs. 50 per order and the
carrying cost fraction is 0.15. All the assumptions of the basic EOQ model are
applicable. The company, however, can take advantage of one of the several
quantity discounts. The pricing schedule of the pistons is :
Quantity Unit Price (Rs)
0-1999 65
2000 - 4999 60
5000 - 10,000 55
above 10,000 50
(a) What is the optimal order quantity?
(b) What is the minimum inventory costs?
Answer : Q* = 10,000 units TVC = Rs. 937590
9. A company purchases a certain chemical to use in its process. No shortages are
allowed. The demand for the chemical is 1,000 %/month; the cost of one
purchase is Rs. 800; and the holding cost for one unit is Rs. 10 per month. The
unit cost depends on the amount purchased as given below :
Amount in kg Cost Per kg (Rs.)
0-249 8
250 - 449 7.50
450 - 649 7
650 & above 6.75
Determine the optimum purchase quantity and the optimum total yearly costs.

Answer: Q* = H2.857 kg TVC = Rs. 19314

1X1 Stochastic or Probabilistic Models of Inventory Control

A class of inventory problems requires that the order quantity decision be made
only once for the entire demand process. And this is also not certain and requires to
make use to probabilities since the decision of dilemma concentrates on ordering
'enough', so that the full potential for profit may be realised, but not too much so as
to avoid losses on the excess.
There are numerous examples in nature, a few are listed below.
1. How many crackers are to be manufactured and stocked for this Deewali season.
2. How many lunch plates are to be prepared for the party.
3. How many leaves of bread are to be baked for a given day
4. How many news papers are to be stocked to sell on a given day.
5. How many salwar suits of current fashions & styles are to be produced.
Inventory Management 401

In all the above examples of stochastic inventory models, the major

distinguishing characteristic is that the order quantity decision is a "one shot" affair
and that, despite uncertainty in the demand, a proper trade-off must be achieved
between the consequences of too much and too little.
Thus in the production systems also, the decision regarding optimum batch
quantity becomes so critical and complicated that we have to depend on the
probabilities to determine the demand first. Thus entire problem depends on the
probable demand again.

Again the use of probabilities is extendable to various distributions such as

Poisson, Normal and other models as applicable to the situation. The stochastic
inventory control decisions also depend on the other factors such as with or with out
set-up cost, instantaneous or continuous demand, descrete and continuous
replenishments, and shortages allowed or not allowed and so forth.

One popular and most often occurring model of probabilistic nature is

discussed here i.e., single period probabilistic model without set-up costs (with
instantaneous demand with discrete replenishment unit, no set-up cost).

The retail marketers, news paper sellers, seasonal products sellers often have
such problem of decision regarding how much stock is to be maintained. This decision
is purely probabilistic and may depend on the past record (forecasting otherwise) and
thus expected (probable) demand during the period.

In such cases, the following inventory model will be helpful.

7.18 Model 7: Single Period Probabilistic Model With

Instantaneous Demand Without Setup Cost
The following working rule is used :
Working Rule :
C —C
Step 1 : Calculate ———
Cc + Cs
Where Cs = shortage cost; C = cost of item per unit (if considered) Cc = carrying
cost

Step 2 : Calculate cumulative probability distribution (for each period).

Step 3 : Determine the value of Q* and ((?* - 1) where the ratio calculated in step - 1
lies in distribution calculated in step - 2.
Step 4 : Take the higher value as optimum level
Step 5 ; if / — inventory on hand before the order is placed, then order Q? -1 if
£>* > / and do not order if Q* < /
402 Operations Research : Theory and Practice
ILLUSTRATION 11

A newspaper boy buys papers for 3paise and sells them for 7paise each. He can
not return unsold newspapers. Daily demand has the following distribution.

No. of
23 24 25 26 27 28 29 30 31 32
customers
Probability 0.01 0.03 0.06 0.10 0.20 0.25 0.15 0.1 0.05 0.05

If each dayfs demand is independent of the previous day's, how many papers
should the order each day? [JNTU (CSE) 97/S]

Solution ;
Let Q, be the number of newspapers ordered per day and let d be the demand
for it, i.e., the number of newspapers actually sold per day.
Let Cj = cost price of newspaper and
C2 = Gain i.e., C1 - sales price
Given C} = Rs. 0.03 and C2 = Rs. 0.07 - Rs. 0.03 = 0.04
C2 0.04 0.04
= 0.57
0.03+0.04 0.07
The optimum solution is obtained by developing the cummutative probability
distribution of daily demand as follows.
d 23 24 25 26 27 28 29 30 31 32
P(d) 0.01 0.03 0.06 0.10 0.20 0.25 0.15 0.10 0.05 0.05
a 0.01 0.04 0.10 0.20 0.40 0.65 0.80 0.90 0.95 1.00
1
r=23

Since = 0.57 is between 0.4 and 0.65, the condition for optimality

suggests that Q°=28 i.e., optimum number of newspapers to be ordered is 28.

Practice Problems
1. A newspaper boy buys paper for 60 paise each and sell them for Rs. 1.40 paise
He cannot return unsold papers. Daily demand has the following distribution.
No. of Customers 23 24 25 26 27 28 29 30 31 32
Probability 0.01 0.03 0.06 0.1 0.2 0.25 0.15 0.1 0.05 0.05
If each day's demand is independent of the previous day's how many papers
should be ordered each day ? [JNTU (Mech.) 97]
Answer: 28
Inventory Management 403

2. A newspaper boy buys papers from Rs. 1.40 and sells them for Rs. 2.45 each.
He cannot return unsold newspapers. Daily demand has the following
distribution:
No. of Customers 25 26 27 28 29 30 31 32 33 34 35 36
Probability 0.03 0.05 0.05 L_0.1 0.15 0.15 0.12 0.1 0.1 0.7 ^0.6 0.02
If each day's demand is independent of the previous day's, how many papers
he should order each day? [Meerut M.Sc. (Math), 1993,1992, JNTU (CSIT) 98]
Answer; 30
3. A contractor of second hand motor trucks uses to maintain a stock of trucks
every month. Demand of the trucks occurs at a relatively constant rate but not
in a constant size. The demand follows the following probability distribution
Demand 0 1 2 3 4 5 6 or above
Probability 0.4 0.24 0.2 0.1 0.05 0.01 0.00
The holding cost of an old truck in stock for one month is Rs. 100 and the
penalty for truck if not supplied on demand is Rs. 1,000. Determine the optimal
size of the stock for the contractor. [Marathwada M.Sc. (App. Math), 1982]
Answer: 3
4. A T.V dealer finds that costs of holding a television in stock for a week is Rs.
20, customers who cannot obtain new television immediately tend to go to another
dealer; and he estimates that for every customer who does not get immediate
delivery he loses on an average Rs. 200. For one particular model of TV, the
probabilities for a demand 0, 1, 2, 3, 4, and 5 in a week are 0.05, 0.1, 0.2, 0.3, 0.2
and 0.15 respectively. How many television per week should the dealer order?
Answer: 4
Hint: CY = 20 C2 = 200, 0.85 < 0.921 < 1.00 [JNTU (CSE) 98/S]

5. The probability distribution of monthly sale of certain items is as follows :

Monthly Sales 0 1 2 3 4 5 6
Probability 0.01 0.06 0.25 0.35 0.2 0.03 0.1

The cost of carrying inventory is Rs. 30 per unit per month and the cost of unit
shortage is Rs. 70 per month. Determine the optimum stock level which
minimise the total expected cost.
Answer : 4
6. Some of the spare parts of a ship costs Rs. 1,00,000 each. These spare parts can
only be ordered together with the ship. If not odered at the time when the ship
is constructed, these parts cannot be available on need. Suppose of loss of Rs.
10,000,000 is suffered from each spare part that is needed when none is available
in stock. Further, suppose that probabilities that the spare part will be needed as
replacement during the life-time of the class of the ship discussed are
Spare part Require 0 1 2 3 4 5 or more
Probability 0.9488 0.04 0.01 0.001 0.0002 0
How many spares should be procured?
Answer: 2
404 Operations Research : Theory and Practice

7. A newspaper boy buys papers for Rs, 2.60 each and sells them for Rs. 3.60 each.
He cannot return unsold newspapers. Daily demand has the following
distribution.
Customers 23 24 25 26 27 28 29 30 31 32
Probability 0.01 0.03 0.06 0.1 0.2 0.25 0.15 0.1 0.05 0.05
If each day's demand is independent of the previous day's how many papers
should he order each day? [JNJU (Mech. 97/S]
Answer: 27

7.19 Inventory Control Approaches

In the previous topics, we have discussed on "how much to order", its related
re-order level and its cost calculations. However, another important point of operating
doctrine of inventory is "when to order?". Though, this depends on many factors and
the reorder point is set basing on several aspects, there are two systems often chosen
by materials managers. These are :
1. P - system or fixed period system
2. Q, - system or fixed quantity system.
1. P - System or Fixed Period System or Periodic Review System :
In this system, the orders are placed at a predetermined fixed period
irrespective of the quantities in stock. However, the re-order is placed for the
quantities that are less to the target (maximum) inventory level. A review is made
periodically and the period is fixed based on the estimated consumptions or past
records. A graphical representation is shown below.
Inventory
Quantity order N cycle
-H*
Maximum
inventory

Reorder
Reorder
level
level \ Reorder period \
Reserve bin

(a) P - System
(b) Q - System
FIGURE 7,20 P&Q-SYSTEMS

2. Q - System or Fixed Quantity System or Two - Bin System :

In this system, a re-order level is fixed- and whenever this level is reached,
irrespective of time (or period), the order will be placed for procurement. Here, the
reorder periods of different cycles are different and inventory cycles also differ since
the re-ordering depends on 'when' the quantities reach the pre-determined minimum
level. This is represented in the graph given above.
Inventory Management 405

An example with a over head water tank, we can understand the above systems
easily. Suppose, you depend on the municipal water supply, which gives you water
only in the morning at 6 am everyday. Then you will fill your tank at 6 am everyday
irrespective of the stock in the tank. This is P - system.
If you have a borewell in your campus, you will fill the tank whenever a minimum
level is reached irrespective of period of consumption. This is £)- system. Now, apply
the above systems for inventory, assuming tank a« your inventory store and water is
items procured.
Some More Examples :
P -System : A government employee who receives salary during the first week
dumps all the grocery (inventory) on every first week of the month.
Q- System: A businessman may not bring his household grocery at every certain
period, but brings whenever a minimum level is reached.
The petrol/diesel in the fuel tank of your scooter /bike or any vehicle has two
bins (though not separate), called main bin and a reserve bin. Whenever, the stock
level reaches reserve, you start your procurement action which is in support of Q -
system, while you may check-up air in the tyres periodically (say weekly) which is in
P - system.
Distinction Between P-system and iJ, - system :
S.No. P - system Q - system
1. This system is based on fixed period or time This system is based on fixed minimum quantity
level.
2. Cycle period and re-order period are Cycle period and reorder period change
constant
3. Cycle period is equal to reorder period. Cycle period is not equal to reorder period
4. Only one bin is used 2-bins (main & reserve) are used.
5. Advantageous for joint production/ Advantageous in situations where stock-out costs are
transportation buying high
6. Useful where financial resources are Useful where financial aresources are abundant
avialable at fixed intervals or/and available at any time.

720 Stores Management & Selective Inventory Control

In general situations of materials management, all the items will not have same
importance due to their cost, consumption, value, function etc,
Thus the care to be taken obviously differs, based on the importance of
materials. To have a control and modeling on the materials which are of less cost or
importance will be uneconomical.
Therefore materials managers often use selective inventory controlling
methods by grouping all the item in certain categories and control them with some
basis.
Various techniques available for managers in this regard are ABC analysis,
VED, SDE, XYZ, SOs, ESND, VEIN, CIN, GOLF etc. Of these, ABC analysis is the most
popular and widely employed technique.
406 Operations Research : Theory and Practice

ABC Analysis :
ABC analysis is a basic tool, which helps materials mangers to take a decision
regarding "how much care is to be put on each item". This selective approach is based an
"annual consumption value" of various items.
For example, the items like nuts, bolts, valves etc., are though equally important
as compared to engines, pistons, cylinders, gears etc., the costs differ. Also the
consumption varies from item to item. The items like engines are stocked in less
quantities due to their high cost and low consumption while high stocks of nuts, bolts
etc., will be maintained as they are cheap and more consumed. Thus more emphasis
should be given to control the stocks of engines while it is expensive and uneconomical
to maintaining and recording the bolts, nuts etc.,
Thus depending the consumption and costs of the items, the items are classified
into three categories A, B and C in the disc ending order of their annual consumption
cost. Thus usually, A - class items are costly but less consumed while C - class items are
many cheap and more consumed. The average pattern of percentage of items and
percentage of annual consumption costs are as follows.
Category Percentage of Items Percentage of Annual
Consumption Cost
A 10%-20% 70% - 85%
B 20% - 30% 10%-25%
C 50% - 70% 5%-15%
ABC (Always Better Control) Technique is in accordance with Pareto's principle
"VITAL FEW- TRIVIAL MANY1 and thus low value items are usually consumed more
and high value items are less consumed.

FIGURE 1.2I : A B C ANALYSIS • GRAPHICAL REPRESENTATION

The computation algorithm for ABC analysis is as follows :

Step 1 ; Obtain data of annual consumption rate and unit cost of each item.
Step 2 ; Calculate annual consumption cost of each item
Annual cosumption cost = annual consumption x unit cost
Step 3 : Arrange the items in the descending order of their annual cost consumption
values.
Step 4 : Find the cumulative value of annual consumption cost.
Step 5 ; Express the annual consumption cost as a percentage of total value of the items.
Step 6 ; Express percent number of items in the total number of items
Inventory Management 407
Step 7: Draw graph between the cumulative precent items to cumulative percent
annual consumption cost, i.e., graph between the values of step - 5 & step 6.
Step 8 : Classify the area under graph into three class A, B and C according certain
norms such as given below.
Class Percentage of Items Percentage of Annual
Consumption Cost
A 10 80
B 20 15
C 70 5

ILLUSTRATION 12
A company is considering a selective inventory using the following data. Classify
them using ABC analysis.
Item 1 2 3 4 5 6 7 8 9 10 11 12
Consumption 6000 61,200 16800 3000 55800 22630 26640 14760 20520 9000 29940 24660
Unit Cost 4.00 0.05 2.10 6.00 0.20 0.50 0.65 0.4 0.4 0.1 0.3 0.5

[OU (MBA) 92]

Item Consumption Unit Cost Consumption Ranking

Cost
1 6000 4.00 24000 II
2 61200 0.05 3060 XII
3 16800 2.10 35280 I
4 3000 6.00 18000 III
5 55800 0.20 11160 VII
6 22680 0.50 11340 VI
7 26640 0.65 17316 IV
8 14760 0.40 5904 XI
9 20520 0.40 8208 X
10 9000 0.1 9000 VIII
11 29940 0.3 8982 IX
12 24660 0.5 12330 V

S.No. Item Annual Comulative % Comulative % Items Commulative

No. Consumption ACC ACC % Items
Cost (ACC)
(1) (2) (3) = Z (2) (4) = (3)/164580 (5) = (1)/12
1 3 35280 35280 2L44 8.33 A
2 1 24000 59280 36,02 16.67 A
3 4 18000 77280 46.96 25.00 B
4 7 17316 94596 57.47 33.33 B
408 Operations Research : Theory and Practice

5 12 12330 106926 64.97 41.67 B

6 6 11340 118266 71.86 50.00 B
7 5 11160 129426 78.64 58.33 C
8 10 9000 138426 84.11 66.67 C
9 11 8982 147408 89.57 75.00 C
10 9 8208 155616 94.55 83.33 C
11 8 5904 161520 98.14 91.67 C
12 2 3060 164580 100.00 100.00 C

10 20 30 40 50 60 70 80 90 100
Percentage item
FIGURE 7 . 2 2 : GRAPHICAL REPRESENTATION OF A-B-C ANALYSIS

F-S-N Classification : This is most widely employed control tool in marketing

and in production. This classification takes into account the pattern of issues from
stores; The three letters stand for Fast vioving, Slow moving and Non moving (some
people refer to D also, which stands for Dead items). This becomes very handy when
it is used to control obsolescence. Items classified as S and N require very great
attention, especially N items. There are several reasons for an item falling in N
category like change in technology, in the specifications of a particular spare part/item
might no longer be in use. However, when a fast moving item is not moved with its
pace, it may go into slow moving and in turn may become non moving also. Therefore
a manager has to be very careful in managing their movement or consumption we
can also observe here that in a store where the items are placed according to the FSN,
the lF items will occupy a forefront places while W items occupy last or backward
place. Therefore a manager has to keenly observe and act at right time to create the
proper demand to these items, else they may become dead or obsolete. When an FSN
classification is made all such information stands out prominently enabling managers
to act on the information in the best interests of the organisation.
V-E-D Classification : This classification is based on the significance of the item
in production or assembly chain, the three letters, V, E, D stand for VITAL,
ESSENTIAL and DESIRABLE items, the items irrespective of their costs that are
highly important without which you cannot run the production or assembly are vital.
The absence of essential items may not stop the production chain but suffers
Inventory Management 409

production chain considerably. The desirably items are usually required to make the
product good looking or aesthetic and their absence hardly has any impact on
production chain. For a scooter, engine is a vital part, breaks are essential and seat
cover is desirable. Note that a small value at the tyre is also vital though cheaply
available.
The other selective inventory control techniques are summarised and
presented in a tabular form as follows.
S.No. Classification Meaning Basic Principle Typical Area of
Application
1. ABC A - High consumption value Percentage of total Common and widely
B - Medium consumption value consumption value employed to any area
C - Low consumption value
2. HML High cost, Medium cost Cost of the items Precious/costly items
and Low cost items
3. VED Vital, Essential and Desirable Item significance in Spare parts/ MRO
the operation or
machine
4. VEIN Vital, Essential, Important and -do- -do-
Normal
5. GOLF Government, Open market, Availability & Items under restrictions or
Local market, Foreign market Restrictions constraints
6. SOs Seasonal, Off seasonal Season of availability Anticipatory inventory
7. FSN/FSND Fast moving, Slow moving, Non Market trend Marketing sales inventory,
moving, Dead finished goods inventory,
store issue patterns.
8. XYZ X - High consumption Consumption General items and
Y - Medium consumption Regularly used items
Z - Least consumption
9. CIN Critical, Important Be Normal Criticality Machinery, Equipment,
Components etc.,
10. SDE Scarece, Difficult, Easy available Availability of the Market demand items.
item in the market

Review Questions
1. Derive EOQ formula, stating the assumption. [JNTU. B.Tech Mech 99]
2. Derive a formula for optimum batch quantity if the demand is continuous at a
constant rate. Assume appropriate data. [JNTU B.Tech CSE 1997]
3. What is ABC analysis in inventory control?
Describe it in detail with the help of graph. [JNTU B.Tech Mech 1998/S]
4. Write a short note on purchasing inventory with one price break.
[JNTU B.Tech Mech. 99]
5. Discuss the significance of stochastic models in inventory control of production
system [JNTU B.Tech CSE 1997]
410 Operations Research : Theory and Practice

6. What are the assumptions used in deriving EOQ formula in Wilson - Haris
model.
7. What assumption do you take for obtaining E O Q for deterministic inventory
models
(a) With shortage
(b) With a production rate
8. Derive the expression for EOQ in deterministic model with back orders.
9. Derive the formula to find EOQ with the production rate (r) and demand {d)
without shortages. How does the formula be interpreted if the same model is
assumed to have shortages. Interprete these with graphically.
10. What are the functions of inventory. [JNTU Mech. 2000/S, CSE - 2000]

11. What are the advantages of having inventory. [JNTU Mech, Mechatronics,
prodn. & Civil 2001]

12. Discuss various costs involved in inventory control in an industry.

[JNTU Mech. 96/MS]

13. Discuss the role of inventory in a large scale industry. [JNTU Mech 96/S/MS

14. What are the consequences of over- inventory and under - inventory situations.
15. Discuss the advantages and disadvantages of allowing shortages in the
inventories. [JNTU CSE 95/S]
16. What is selective inventory control. Why do you optimise this in large industries.
[JNTU Mech. 96/PPC]
17. What are P - system 8c Q - system of inventories ? Explain.
[JNTU Mech. 95/S, 98, 99/PPC]
18. Distinguish between P - system 8c Q, - system of inventories.
[JNTU Mech. 95, 96, 96/S, 97/PPC]

19. Discuss the working rule of probabilistic model without set-up cost and
instantaneous demand.
20. Give the algorithm for price break models
21. Write the algorithm for ABC analysis. [JNTU - Mech. 98/S/PPC]
22. Discuss the factors affecting inventory in the industries. [JNTU Mech. 93/MS]
23. Write short notes on
(a) Decoupling inventory
(b) Transit inventory
(c) Buffer inventory [JNTU Mech. 93/0R
24. Give the classification of inventories.
25. Classify and discuss various inventory models briefly. [JNTU - ECE/EEE - 93/MS]
26. Discuss various selective inventory control techniques and compare them.
Inventory Management 411

27. Derive the EOQ formula for continuous & non uniform demand with
instantaneous procurement.
28. Write short notes on
(a) Lead time (b) Inventory cycle
(c) Re-order cycle (d) Re-order period. [JNTU Met & CSE - 93/S/MS]
29. What do you understand by the term operating doctrine in the inventory
modeling.
30. What are the limitations of EOQ.

Objective Type Questions

1. At EOQ
(a) annual purchase cost = annual ordering cost
(b) annual ordering cost = annual carrying cost
(c) annual carrying cost = annual shortage cost
(d) annual shortage cost = annual purchase cost

2. If shortage cost is infinity

(a) order cost is zero (b) no shortages are allowed
(c) no inventory carrying is allowed (d) purchase cost = carrying cost

3. The stock maintained to withstand unknown demand changes is known as

(a) De-coupling inventory (b) pipeline inventory
(c) fluctuatory inventory (d) anticipatory inventory
4. The inventory policy followed for water consumption from a water tank
operated with municipal water spply connection is usually
(a) P - system (b) Q - system
(c) 2 - bin system (d) Stochastic model
5. The inventory system designed for a petrol tank of a two-wheeler is
(a) P • system (b) Q - system
(c) Stochastic system (d) Poisson model
6. The most suitable analysis for a retail marketer is
(a) F - S - N analysis (b) A - B - C analysis
(c) G - O - L - F analysis (d) V - E - D analysis
7. Which of the following inventory is maintained to meet expected demand
fluctuations
(a) buffer stock (b) de-coupling inventory
(c) fluctuatory inventory (d) anticipators inventory
412 Operations Research : Theory and Practice

8. With usual notations the formula for EOQ when shortage is allowed and finite
and uniform supply rate

(b)
-d

2C 0 Z)

9. Expression for total variable cost in the case of instant replenishment without
shortage, with usual notations.
(a) VC0 Cc D (b) VC0 Cc D/2 (c) V2 Co Cc D (d) none of these
10. If EOQ = demand (D), then the number of orders placed in the inventory
horizon (i.e., 1 year ) is
(a) zero (b) one (c) two (d) infinite
11. If average carrying cost per unit per year are twice to that of ordering cost/order
then EOQ varies with
(a)Z)2 (d)D (c)VD (d)V2D
12. Being ordering cost/order and average unit carrying costs constant, if demand
suddenly falls by 75% then EOQ
(a) increases by 50% (b) decreases by 50%
(c) decreases by 37.5% (d) none of these
13. Which of the following increases with quantity ordered per order
(a) order cost (b) carrying cost
(c) demand (d) purchase cost
14. The time difference between ordering point and replenishment point is called

(a) re-order period (b) lead time

(c) inventory cycle time (d) inventory horizon
15. Which of the following is assumed to be constant in deterministic inventory
models
(a) annual demand (b) annual ordering cost
(c) annual carrying cost (d) annual shortage cost
16. When shortages are allowed with constant demand and instaneous
replenishment, EOQ = (with usual notations)

(b)

2C0D Cc-Cs
y ——
L.
• —— (d) y
c
 413
Inventory Management

17. Shortages are allowed in an industry to derive the advantage of

(a) spreading the ordering cost over a longer period

(b) less number of orders in the inventory horizon that reduces ordering cost
(c) increased carrying cost does not effect considerably
(d) all the above

18. In the case of shortages, the optimal amount of back order units = _ ^

(a)^=p (b) 10% of EOQ

C C
(c) EOQ . — - V (d) EOQ . — c —

19. Total cycle time in case of model with shortages is

Cs

I L' ) cc + cs

20. A continuous and finite supply rate model becomes simple EOQ [i.e., instant
supply] model for the condition that production rate =

(a) zero (b) unity

(c) infinity (d) consumption rate
21 - Total variable cost becomes zero if

(a) rate of production is infinity

(b) rate of consumption is infinity
(c) rate of production = rate of consumption
(d) never
22. The optimal number of shortages in the inventory model with finite and
constant supply with shortages allowed is (with usual notations)

(d)
r-
414 Operations Research : Theory and Practice

23. In the model with continuous production with shortages, the production time
is
(a) EOQ/rate of produce (b) EOQ/rate of consumption
(c) EOQ'annual demand (d) none of the above
24. No. of orders in year =
(a) annual demand/total cost (b) annual demand/EOQ
(c) EOQ/demand (d) EOQ/rate of production
25. The optimal stock level of a shortage model (with instantaneous procurement)
is

I2QD

c $
•V^^
. /2C.Z)
c
C-
c

26. The annual demand of an item is 1200 units and placing the order costs Rs. 12
per order and costs 20% of price to carry. The price of the item is Rs. 10/- The
EOQ =
(a) 12 (b) 120
(c) 1200 (d)20
27. The number of orders to be placed in the year in the above problem is
(a) 12 (b) 10
(c) 6 (d) 100
28. The re-order period for the above problem (Q. No. 26 ) is
(a) monthly (b) quarterly
(c) half yearly (d) yearly

29. Which of the following is the inventory purchased to get the advantage of price
discounts or to reduce transportation costs
(a) fluctuatory inventory (b) lot wise inventory
(c) anticipatory inventory (d) transit inventory

30. The material stored to safeguard the production chain in the event of machine
breakdown is termed as __
(a) transit inventory (b) lot wise inventory
(c) buffer stock (d) decoupling inventory
Inventory Management 415

31 Operating Doctrine of inventory is concerned with

(a)when to order and how to order

(b) when to order and who to order
(c) when to order and how much to order
(d) who to order and how to order

32. The purchase cost is assumed to be varying with quantity ordered in the
inventory model with

(a) shortages (b) finite production rate

(c) price breaks (d) none

33. In which case of the following, probabilities inventory models are not
employed

(a) stocks of retailers (b) news papers seller

(c) dealer of single commodity (d) a manufacturer of crackers

34. The optimum stock is determined in probabilities model with instant demand,
no set-up cost, where the cumulative probability just exceeds the ratio
of
(a) shortage cost to sum of shortage and holding costs
(b) holding cost to sum of shortage and holding cost
(c) sum of shortage and holding cost to shortage cost
(d) shortage cost to difference of holding and shortage cost.

35. Stochastic model means

(a) deterministic models (b) probabilistic models

(c) fixed period models (d) fixed quantity models

36. Which inventory costs are assumed to be zero in JIT inventory system
(a) ordering cost (b) transportation cost
(c) carrying cost (d) purchase cost

37. Taxes and insurance are included in

(a) ordering cost (b) shortage costs

(c) purchase cost (d) carrying costs
416 Operations Research : Theory and Practice

38. Which of the following is false in the concept of EOQ and its assumptions
(a) lead time is zero
(b) annual order cost = annual carrying cost
(c) ordering cost remains same through out the year
(d) none of the above

39. The criteria used inA- B - C Analysis is

(a) cost (b) value
(c) consumption (d) consumption cost
40. Which of the following analysis neither considers cost nor value
(a)A-B-C (b)X-K-Z
(c)H-M-L (d)V-E-D

41. The inventory controlling technique that uses significance of the item as the
basis is
(3)A-B-C (b)F-S-N
(c)V~E~D (d)H~M-L

42. Items sold under the restrictions such as government policies, FERA, MRTP
act etc, are classified by
(a) ABC (b) FSND
(c) VEIN (d) GOLF

43. The availability of item in the market is basis for analysis

(a) ABC (b)SDE
(c) SOs (d) HML
44. S-D-E analysis means
(a) source - destination - EOQ (b) single-double-empty
(c) scarce - difficult-easy (d) specific - desirable-essential

45. If average unit carrying cost is equal to order cost per order and demand is 200
units/year then EOQ is
(a) 20 (b) 40
(c) 10 V§^ (d) insufficient information
46. When Cc is twice Co to make EOQ = 20, demand must be

(a) 200 (b) 400

(c) 100 (d)20
Inventory Management 417

47. Simple EOQ is independent of

(a) holding cost (b) order cost

(c) shortage cost (d) purchase cost

48. If demand is equal to EOQ, then

(a) single Order will be placed per year

(b) order cost per order is equal to annual carrying cost
(c) order cost per order is equal to half of total variable cost
(d) all the above

49. If annual demand is square of optimal order quantity, then

(a)C 0 = Cc2 (b)C0 = 2-Ce

{c)Cc = d (dK=2C0

50. In F-S-N analysis of inventory control, '5* represents

(a) scarcely available items (b) supply rate uniform
(c) slow moving items (d) seasonal items

51. Which of the following product is an example for seasonal consumption but
continuous production

(a) sugar (b) electricity

(c) rain coats (d) news paper

52. Which of the following product is not an example of seasonal production but
continuous consumption

(a) sugar (b) mango

(c) rain coasts (d) soaps

53. The pair of inventory models with same EOQ formula, when total inventory
period is one year is

(a) instant procurement with shortage & without shortage

(b). finite & continuous supply rate and instant supply
(c) simple EOQ with uniform demand rate and different rates of demand in
different cycles
(d) shortage model and no set-up cost and probabilistic model
418 Operations Research : Theory and Practice

54. The total inventory period is referred to as _ ^

(a) cycle (b) horizon
(c) ROP (d) Lead time
55. Costs for tenders and auctions are included in
(a) order cost (b) holding cost
(c) shortage cost (d) purchase cost

56. Which of the following is considered as as a demand independent item

(a) a wrist watch (b) a glass on wrist watch
(c) a jewel in watch (d) none of the above

Fill in the Blanks

1. An inventory model with shortages will become simple EOQ model if shortage
cost is
2. The EOQ formula was first proposed by
3. The stock maintained to encounter the production disturbance due to machine
breakdown is
4, A-B-C analysis is based on
5. If a single order is placed in the total inventory period, then EOQ =
6. The total period for which inventory decision are made (usually one year) is
called
7. The decisions such as when and how much to procure are said to be
of inventory
8. For production rate = , the model with finite production rate
becomes simple EOQ model
9. The order cost/order with the increase in ordered quantity
10. The time difference between any two consecutive orders is called
11. The time difference between any two successive replenishments is called
12. Keeping Co and Cc constant, if demand doubles suddenly then EOQ
13. In case of shortages, the ratio of optimal amount of back order to optimal order
quantity is equal to
Inventory Management 419

14. Inventory cycle time in the case of finite and uniform supply rate is .

15. The optimum number of orders placed = .

16. If CQ is half of Cc, then annual demand is of EOQ

17. F-S-N-D analysis is abbreviated for F = _ _ _ _ _ ^ _ S = _^__

N= D= •

18. The inventory controlling technique used on the basis of cost of annual
consumption is
19. Formula for total costs in simple EOQ model is

20. The time gap between the ordering point and its immediate actual stock
completion point is called

21. Fill in the boxes :

ATC= «x—§ • +• x^+ +

X
2
xD

22. The average quantity consumed when quantity '£)' is replenished in a cycle of
time '£' is,

23. When an inventory model is shown graphically as time on X - axis, and quantity
on Y - axis, the area under the graph represents .
24. Given commulative probability is 0.25, 0.5, 0.75 for sale of 2, 3 and 4 units of
stock respectively with storage cost Rs. 2.50 ps and shortage cost Re. 1/-, the
optimum number of units to keep in stock is

25. Given order cost Rs. 20 per order, annual demand = 1000 units, carrying cost
10% of unit price Rs. 10/- of an item, the EOQ = ,

26. The no. of orders in the above question =

27. If there are 300 working days, the ROP in the above question is =

28. The operating doctrine of inventory models is to find and to

order.
29. The quantities kept in stock to safe guard the product in the event of excess
consumption or unusual demand fluctuations etc. is called .
30. The items whose demand is estimated based on the demand of the final product
to which it is assembled is called
420 Operations Research : Theory and Practice

Answers
Objective Type Questions :
Mb) 2. (b) 3. (c) 4. (a) 5.(b)
•6: (a) 7.(d) 8. (d) 9.(c) 10. (b)
11. (c) 12. (b) 13. (a) 14. (b) 15. (a)
16. (c) 17. (d) 18. (d) 19. (d) 20. (c)
21. (c) 22. (b) 23. (a) 24. (b) 25. (b)
26. (b) 27. (b) 28. (a) 29. (b) 30. (d)
31.(c) 32. (c) 33. (c) 34. (a) 35. (b)
36. (c) 37. (d) 38. (d) 39. (d) 40. (d)
41. (c) 42. (d) 43. (b) 44. (c) 45. (a)
46. (b) 47. (d) 48. (d) 49. (d) 50. (c)
51. (c) 52. (c) 53. (c) 54. (b) 55. (a)
56. (a)
Fill in the Blanks :
1. infinity 2. Wilson & Harris
3. decoupling inventory 4. consumption cost
5. demand 6. inventory horizon
7. operating doctrine 8. infinity
9. decreases 10. re-order period or ROP
11. inventory cycle period 12. increases by V2~ times

is ^
14. EOQ/consumption rate

, r Demand 16. square

EOQ
17. fast moving, slow moving, non moving, 18: ABC analysis
dead
19. V2Co, CeD +f D 20. Lead time
21. order cost/order (Co), carrying cost on avg.
qty. (Cc) unit price (p) quantity l(? 22. f
23. consumption quantity in the period 24. 3 units
25. 200 units 26. 5 orders
27. 60 days 28. when and how much
29r buffer or reserve stock 30. demand dependent item
422 Operations Research : Theory and Practice

CHAPTER AT A GLANCE
patjsflBsstfl

JU3ISUBJ1 - pagspes

pagspssuQ

(a) wn

Hi
CO

(Odli) TO Wii m jsei

(Oili) TO wn J °1 ^ F i

1
Queue Models (Waiting Lines) 423

8.0 Introduction
One of the common situations in our daily life occurs with waiting or queue
system. Rather, there will not be any person on this earth who did not experience the
troubles due to waiting such as at bus stops, ticket booths, doctor's clinic, petrol bunks,
bank counters, traffic lights and so on. Queues are very common in industries also, in
shops where the machines wait for repair/maintenance, operators wait at tool cribs,
material waiting for operations etc. These queues ruin the time of customers and it
may even cost hugely in some occasions. The minimisation of waiting time and thus
costs are the primary concerns of these queue models in Operation Research.

8.1 Terminology
The following terms are commonly used in queue models.
1. Customers : The persons or objects that require certain service are called
customers.
2. Server : The person or an object or a machine that provides certain defined
service is known as server.
3. Service : The activity between server and customer is called service that
consumes some time.
4. Queue or Waiting Line : A systematic or a disciplined arrangement of a group
of persons or objects that wait for a service is called queue or waiting line.
5. Arrival: The process of customers coming towards service facility or server to
receive certain service is Arrival.
6. Mean Rate of Arrival : The average number of customers arriving per unit
time is called mean rate of arrival and is denoted by T ( read as LAMBDA).
X = Total no. of customers arriving/total time taken.
7. Mean Inter-arrival Time : It is the average time gap between two consecutive
arrivals of customers. It is the inverse of mean arrival rate i.e.,
Total arrival time
Number of arrivals
8. Rate of Service : It is the average number of customers served per unit time
and is denoted by (X ( read as MUE).
Total number of customers served
Total service time
9. Mean Service Time: It is the average time taken by the server to serve a customers
,. , . Total service time
and is equal to inverse ofn service rate i.e., 1/u = — 7.
Number or customers served
424 Operations Research : Theory and Practice

8.2 When Does A Queue Result?

Suppose a queue system whose arrival rate is X and service rate is |x.

Arrival rate Service rate

(X) (M)

Arrivals Service Served customers

facility

F I G U R E 8 . 1 - : QUEUE OR WAITING LINE

Now, in this system any of the following three relations must occur (by law of
trichotomy).
(i) X > (i (ii) ^ = ^i (iii) X<[i
(i) When X>\it i.e., the rate of arrival is greater than rate of service, i.e., number
of customers arriving is greater than number of customers served per unit time
(time kept constant)^ the system results in piling up of arrived customers waiting
for service and thus a queue results.
(ii) If X = \i, i.e., number of customers arrived is equal to customers served, there
will not be any waiting, and this queue system has cent percent efficiency.
(iii) If X < \L, i.e., number of customers arriving is less than number of customers
served per unit time. This means that server will be free, waiting for customers
to arrive. Thus the queue results in server's point of view.
Thus queue results where there, is an imperfect matching between the rate of arrival and
the rate of service,

8,3 Costs Associated with Waiting Lines (Queuing)

Queuing theory, can be applied to variety of operational situations where there
is an imperfect matching between the rate of arrivals and the rate of service. The rate
of arrival and service however, depend on the probability studies and the imperfect
matching usually occurs due to inability of server to predict accurately the arrival and
service time. These may cost on customer due to waiting or on server to operate the
service facility. Thus the level of service (either service rate or the number of service
facilities) is determined by cost analysis of the following two conflicting costs.
1. Cost of offering service.
2. Cost incurred due to delay in offering service or due to keeping customer
waiting.
Now, we can analyse these as follows.
1. Cost of Service : If existing service level in terms of rate of service or number
of service facilities (channels) is reduced, the cost of operation comes down.
In other words, increase in level of service increases the cost of operating service
facilities.
Queue Models (Waiting Lines) 425

Waiting Costs : In contrast to the above, the reduction in existing service level
will hike the customers waiting time or may result in long queues. Thus the
increase in service level will decrease the waiting costs and decrease in service
level will increase the waiting costs.
Thus the optimum level of service can be obtained by achieving a balance
between the above two costs or at the level where the total cost shows minimum.
This is interpreted through a graph shown below.

Service cost
(to service)

Minimum total cost

Optimum level
of service

Level of service
FIGURE $ 2 : COST ANALYSIS OF QUEUE SYSTEM

8.4 Queue System: Salient Features

A queue system may be studied by diving into two subsystems viz arrival system
and service system. The arrival system is based on customer's behaviours and their
process of arrival while the service system is dependent on the service mechanism and
server's behaviour.
These are diagrammatically shown in the block diagram given at the beginning of this
chapter.
The essential features shown in the block diagram are discussed here below in
detail.
1. Input Source or Population : The input source or population is a set that
contains the probable potential customers to come out for service. Thus the
input source need not be homogenous and may consist of sub populations. The
input source may contain the subset of patients for the queue system at a
doctor's clinic, a sub set of vehicle owners for queue system of petrol bunk, a
subset of passengers who may enter a waiting line for a bus and so on. Another
point to be noted here is that all the population need not be customers, but as
they come out of population with an intention to get the defined service are
considered as the customers to that service system.
2. Arrival Pattern: The customers are expected to arrive at their own convenience
and conditions. However, the time gaps between any two consecutive arrivals,
if noticed, will be fit in certain fashion or pattern. These patterns of arrival times
often follow one of the following probability distribution.
(i) Poisson distribution (represented by M )
(ii) Exponential distribution (represented by M)
426 Operations Research : Theory and Practice

(iii) Erlang distribution (represented by Er )

(iv) General pattern (represented by G)
3. Limit of Queue : (Restricted/unrestricted queues) The limit of queue is some
times restricted by server's behaviour. If the limit is imposed by the server, the
arrivals will be limited for a given period. For example, if a doctor has condition
that he would treat 50 out-patients only per day, the queue will be limited or
finite and 51st out-patient will have to be in queue to enter the queue i.e., he
will be in secondary queue while the first 50 will be in primary queue. However,
if there is no restriction, the queue is said to be unlimited or infinite. The limited
queue is denoted by N while unlimited queue by °°.
Limited queue may be adopted by space restrictions or other conditions also.
For instance, a service facility such as a cinema hall, has restricted
accommodation which can not accommodate more than its space. In such cases
the server will not allow further arrivals unless some space is vacated to
accommodate the new entry. In some other cases, the limited queue system may
be adopted to prevent a formal queue and to avoid nuisances. These queues
are often referred to as 'take-a-number' or 'take-appointment' or 'reservation'
policy. However, the queue may be unlimited for getting into this 'reservation'.
4. Queue Discipline : The queue discipline is the order in which the customer is
selected from the queue for service. There are numerous ways in which
customers in queue can be served of which some are listed below.
(a) First In First Out (FIFO) or First Come First Serve (FCFS) : It is the
discipline in which the customers are served in the chronological order
of their arrivals.
e.g. Tickets at a cinema hall; sales at a grocery shop, trains on a (single line)
platform etc.
(b) Last In First Out (LIFO) or Last Come First Serve (LCFS) : If the service
is made in opposite order of arrivals of customers, i.e., who ever comes
last is served first and first obviously goes to last, it is called LIFO or LCFS
system.
e.g. Stack of plates; loading and unloading a truck or go-down; office filing of
papers in chronological orders; wearing socks and shoes; dressing a shirt and
coat over it, packing systems etc.
(c) Service In Random Order (SIRO) : By this rule, the customer for service
is picked up at random, irrespective of their arrivals.
e.g. Lottery system from which one is picked up, the dresses waiting in a ward robe
from which one is to be chosen, food stuffs in a buffet, sales counter of
commodities or vegetables etc.
(d) Priority Service : Under this rule, the server gives priority to certain
customer(s) due to some importance or prestigeous or high cost group
of the customers.
e.g. A telephone urgent call given to a customer is charged at higher price; a
separate counterfor cheques at a electricity bill payment counter; priority given
at (APSRTC) reservation counter who buys a CAT(Concessinional Annual
Ticket) card etc.
Queue Models (Waiting Lines) 427

(e) Pre-emptive Priority Rule : Under this rule, highest priority is given to
certain customer(s) irrespective of their arrival and costs. The customer
is allowed to enter into service immediately after entering into queue
system, even if another (lower priority) customer is already in service,
i.e., the lower priority customer who is in service will be interrupted
(pre-empted) to facilitate the special customer.
e.g. An emergency case arriving at a doctor's clinic who is attending to a regular
out-patient. (The doctor will stop his service to the regular patient and
immediately rushes to emergency case,)
A minister or VIP coming to receive a service at counter is given highest
priority,
An expediting in production shop.
(f) Non Pre-emptive Priority Rule: This is also a rule by priority to the special
customer but the priority will not emptivate the current service. The
service to the special customer starts immediately after the completion
of current service.
e.g. A medical representative will be given appointment immediately after the
current service to an out-patient at a doctor's clinic.
In a production shop, backlog planning, chase planning is in accordance with
non pre-emptive in many occasions.

Note : In the above disciplines (a) and (b) are called static queue disciplines while the
rest are dynamic queue disciplines.

5. Customer's Behaviour: A customer coming out of population with an intention

to receive certain service may be prepared to wait till he gets service. If he waits
hoping to get the service, he is called 'patient customer' or 'positive customer'
or 'optimistic customer'. Where as a customer may walk out of the queue without
getting service due to various reasons such as long queue in front of him or not
having patience to wait etc., he is said to be an 'impatient customer' or 'negative
customer' or 'pessimistic customer'. Various types of negative behaviours are
defined here below :

(a) Balking: A customer who gets discouraged by seeing the length of the
queue before him and thinks that he may not get service, may walk out
or may not join the queue. He is said to be 'Balking'.

(b) Reneging: A customer who joins the queue and waits for some time but
leaves the queue due to intolerable delay or impatient to wait any longer
is said to be 'Reneging*.

(c) Jockeying; A customer who moves from one queue to another hoping to
receive a more quick service is said to be Jockeying'.
Thus, the customers who come out of their original queue showing a
negative behaviour that he may not get service. However, a customer who
receives service also may have negative bahaviour as given below :
428 Operations Research : Theory and Practice
(d) Unsatisfied Customer ; A customer who is not satisfied by the service by
the quantity is said to be unsatisfied. For example, a customer of an hotel
may not be satisfied with the quantity of food supplied to him though it
is tasty.
(e) Dissatisfied Customer: A customer who is not satisfied by the quality of
service is said to be dissatisfied. For instance, a customer getting tasteless
food at an hotel though supplied in huge quantities is dissatisfied.
The customers unsatisfied or dissatisfied usually expect certain level of
service and when they do not have the actual, equivalent to their
expectation they will not have satisfaction. However, if a customer has
the match between his expectation and the actual receipt, he will be
satisfied and if the actual is more than his expectation, he will be 'delighted'.

6. Type of Service : The service system may be in two ways as given below :
(a) Customer stationary and server moving.
e.g. (i) Arranged meal at which server brings what customer desires,
(ii) A machine waiting for repair.

(b) Server stationary and customer moving.

e.g. (i) A buffet meal where customer goes to buffet table and gets what he wants,
(ii) An aeroplane waiting for a run way for landing.

7. Service Patterns : Similar to that of inter arrival times, the service time taken
by the server to serve each customers varies. These variations depend on the
rate of arrival and server's behaviour. However, these times can be fit into one
of the following probability distribution patterns.
(a) Poisson distribution (denoted by M)
(b) Exponential distribution (denoted by M)
(c) Erlang distribution (denoted by Er)
(d) General fashion (denoted by G)

8. Number of Servers : We have already learnt that a queue system and its
associated costs depend on the level of service which includes both the rate of
service and number of service facilities. On one hand the rate of service is
dependent on the service time distribution patterns while on the other hand
the level of service can be manipulated by arranging number of servers. There
are two ways, of assumption to the queue problem with regard to number of
service facilities.
(a) Single server system represented by 1.
(b) Multiserver system represented by S.

9. Service Configurations : The efficiency and capacity of a queue system can be

increased by arranging the service systems in an effective way, particularly when
multi-service facilities are available. Various arrangements of service facilities
can be classified into following three ways.
Queue Models (Waiting Lines) 429

(a) Series Configuration ;

Arrival Service facilities Departure

Fuel Cash]
Customers V2 service V2 service Served
customers
FIGURE 8 . 3 : SERIES CONFIGURATION
Here service is divided into certain parts served sequentially
(b) Parallel Configuration :
Service facilities
Arrival Departure
Cash
©]_
Fuel I
Fuel and Cash
Full service

Served
Cash
customers
Fuel I
Fuel and Cash
Full service

FIGURE 8 . 4 : PARALLEL CONFIGURATION

Here arrivals are divided into parts and separated in to different queues.
(c) Combined Configuration : This configuration is a combination of the
above two configuration.

Input

Parallel (S, v S2)

FIGURE 8.5 : COMBINATION CONFIGURATION

10. State of Service : We have learnt that the service system is influenced by service
time distribution and on server's behaviour. These are inter dependent on
arrival patterns and customer behaviour. On close observation of the changes
that occur in service system, we can find three states of service viz.,
dy state. (b) Transient state. (c) Steady state.
Unsteady Transient Steady
state state state

FIGURE 8..6: QUEUE STATES

430 Operations Research : Theory and Practice

When the service system is just started, the server will not be able estimate how
much time he has to take per customer. Also, he may be in a bit of confusion and
non-uniformity due to initial conditions and hence takes more time for service. This
state in which the server does non-uniform service is unsteady state.
From this state he slowly moves towards attaining uniformity. This consumes
some time and the server during this time is said to be in transient state.
After getting enough hands on experience, the server picks up and acquires
complete knowledge on the service system and takes almost same times at uniform
rate to serve each customer. This state is called steady state. Further, this steady state
can exist when the ratio of arrival rate to service rate is less than unity.
Steady state condition : — < 1
M-
11. Departure and Output: A served customer goes out of the service zone and
emerges as output. This activity is often known as departure in queuing
terminology. A customer departed from the service zone may have been
satisfied or unsatisfied or dissatisfied. This has already been explained in fifth
point of this section. In the above eleven features, first five are included in
customer zone while the rest in server zone.

8.5 Queuing Models: Kendall - Lee Notations

Queuing models can be defined by five most significant features of the above
explained eleven points. D.G. Kendall in 1953 has noticed three of them in the form
(a/b/c) and later A.M Lee in 1966 added two more in the form (d/e) to describe a queue
model. These are known as Kendall-Lee notations in the standard format as (a/b/c): (d/e)
where
Poisson (M)
Exponential (Af)
a = Arrival distribution
Erlang (Er)
General (G)
Poisson (Af)
Exponential (Af)
b = Service time distribution
Erlang (Er)
Kendall - Lee General (G)
notations for Single (1)
queue model c = No. of service channels (servers)
Multi (S = 2, 3, . .
Limited (N)
d = Limit of customers
Unlimited («>)
FIFO
LIFO
e = Queue discipline SIRO
Pre-emptive
Non pre-emptive
Queue Models (Waiting Lines) 431

A queue model is blend of the above five points, which is diagramatically shown
below.

FIGURES.7: QUEUE MODEL: KENDALL LEE NOTATIONS

Thus if queue system is composed of Poisson arrival distribution, exponential

service time distribution (both of these follow 'Markovian' property and hence
represented by M), single channel, unlimited number of customers in queue can enter
and follow a discipline, First Come First Serve, the queue is described as
(M/M/l) :(-/FIFO)
Similarly (M/Er/s) : (N/SIRO) means Poisson arrivals, Erlangian service time
distribution, s number of channels, limited number of customers are allowed, selected
in random order.
432 Operations Research : Theory and Practice

8.6 Parameters of Queue System

Every queue system has two parameters upon which entire queue system
depends. These are
X = mean arrival rate (or 1/X = mean inter arrival time)
|i = mean service rate (or l/(x = mean service time)

8.7 Axioms ofPoisson Process

The arrivals in a Poisson's process are characterised by the following three
axioms.
Axiom 1: In a non-overlapping interval, the number of arrivals are statistically
independent. This means Poisson has independent increments.
Axiom 2: The probability of occuring more than one arrival between the time
interval, t and [7 + An, is 0 (A t\ This means that the probability of occuring two or
more arrivals during^the smalltime interval At is negligible. Hence we have
PQ (At) + Pj (AT) + 0 (At) = 1
Axiom 3: The probability that an arrival occurs between time, t, and time
(t + AQ, is equal to X (At) + 0(At).
Hence, we have, Px(At) = X(At) + 0 (At)
where X = a constant and independent ofN(t)
N(t) = total number of arrivals upto time t
N(Q) =0

At-*0 At

8.8 Operating Characteristics of Queue System

Some of the operational characteristics, which are of general interest to estimate
the performance of the existing queue system or/and to design a new queue system
are listed below:
1. Expected number of customers in queue or size (or length) of queue (L ) :
This is the average number of customers waiting for service and is denoted by L .
2. Expected number of customers in system or size (or length) of the system
(Ls) :It is the average number of customers expected to be waiting in queue and
being in service. It is denoted by Ls.
3. Expected waiting time in queue (W ): This is average time spent by a customer
in queue before the commencement of the service. This is denoted by W .
4. Expected Waiting time in the System (Ws) : It is the total time spent by a
customer in the system, whcih includes the waiting time in queue and the service
time. It is represented with Ws.

WS = W + mean service time =W +—.

Queue Models (Waiting Lines) 433

5. The server utilisation factor or fraction of busy period or traffic intensity :

It is the proportion of the time that a server actually spends with the customers.
In other words it is the probable period of server being busy. It is designated
with a greek letter p (rho).
Average service completion time (l/|x) X
Average inter arrival time (l/X) \i
General Relationship Between Operating Characteristics :
1. Between L and Ls
Ls = No. of customers in the system
= No. of customers in queue + number of customers in service.
T T AT

2. Between Wq and Ws
Ws = Waiting time in system.
= Waiting time in queue + waiting time in service.

3. Between L and W
Lq No. of customers in queue
W Waiting time in queue
No. of arrivals into queue ^
= X i.e.,• :
— :———
Total time of arrivals into queue I

4. Between Ls and W$ :Similar to the above

5. Expected number of customers served per busy period.

0
P(n>s) p.-X
s - b. rc_ -^_^

= ^ r . p for (M/M/l): (-/FIFO)

\i — A

6. Expected waiting time during busy period,

W. = —2— = ^— for (M/M/l): («»/FIFO).

1 n- X ndi-X)
434 Operations Research : Theory and Practice

8.9 Model -1 :(M/M/1) : (oo/FIFO)

The following expressions are used characterise this model.

1. Probability Density Function Pn = f - . 1- -

= p^(l - p ) where p< 1; n = 0, 1, 2 . . . .
2. Probability of system being empty (i.e., no customer waiting)

3. Probability of server being busy (Pb) = 1 - Po -

4. Expected size of system

5. Expected length of queue

q
X
| fa- X) 1-p
6. Expected waiting time in queue

7. Expected waiting time in system

1 Ls 1
w = Wq +
* ~\i= ~x = \i~x
8. Expected waiting time in the queue for busy system
W W l
Wb = L=
ib TrFo = ^ x = w>
9. Expected number of customers served per busy period.
L
. s 4 U 1
Ph \-Po n-X 1-p
10. Probability that k or more customers waiting in the system.
fx^ fA.f +1
i.e., P(n>k)= - and P(n>k) = \-\
11. Probability of waiting '£ minutes or more in queue
oo

P(W>t) =j-([i-X)e~^
t

ILLUSTRATION 1 —
A, cheques are cashed at a single 'Tellerr counter. Customers arrive, at the
counter in a Poisson manner at an average rate of 30 customers per hour. The
Teller takes, on an average a minute and a half to cash cheque. The service time
has been shown to be exponentially distributed.
Queue Models (Waiting Lines) 435

(i) Calculate the % of time the Teller is busy and

(ii) Also calculate the average time a customer is expected to wait
[JNTU (Mech.) 98/S]
Solution :
In the given problem, arrival process is in Poisson manner (M), service time
distribution is in exponential process (M), No. of servers is 1, No. of customers allowed
in queue process is «>, and the service discipline is FCFS.
Hence the queue model is (M/M/l) : (~ / FCFS) (according to Kendall - Lee
notation).
Arrival rate ( X) = 30 customers/hr.
. (I) ,1 , 3 .
Serivce time — = 1 r mm = — min
hi I 2 2 "
3 1. 1.
= - x - h r s . = - hours.
Service rate ( u.) = 4 0 cusiomers/hr.
X < \x, hence the model-I implies at steady state

(i) Traffic intensity (p)= —

30 3
(or) factor of busy period = — = j = 0.75

The percentage of the time the Teller is busy is

0.75x100 = 75%

(ii) Average length of queue ( Lq) = ^ y or y ^ - = 4Q_SQ =3

Average waiting time that a customer is expected to wait

^ ^ i f ^ s o ' T o ^ 0 ' 1 hours = 6 niin'

ILLUSTRATION 2 —
The mean arrival rate to a service centre is 3per hour. The mean service time is
10 minutes. Assuming Poisson arrival rate and exponential servicing time,
(i) Utilisation factor.
(ii) Probability oftwo units in the system.
(iii) Expected number of units in the system.
(iv) Expected number of units in the queue.
(v) Expected time in minutes, customer has to wait in the system.

Solution :
In the above problem let us first identify the queue system.
Arrival pattern is Poisson distribution i.e., M
Service pattern is exponentially distributed i.e., M
Number of service centres is 1.
436 Operations Research : Theory and Practice

Limit of customers arrival (No limit is imposed) = «>.

Queue discipline = FIFO (When there is no mention of disciple, it may assumed
to be FIFO)
Thus the above system can be described as (M/M/l) : (°°/FIFO) according to
Kendall-Lee.
Next, parameters of the system are to be. identified.
Mean rate of arrival = X = 3 per hour.
Mean service time — = 10 Min = —— = — hrs.
\\L\ 60 6
(X = 6 per hour.
X 3 1
Traffic intensity (p ) = — = — = —

As X/\i < 1, steady state can exist.

Now, we can find all the operating characteristics that govern the queue system.
(i) Utilisation factor = p = — = — = —
\x b 4
2 2 2
(ii) Probability of two units in the system P (n = 2) = p 2 - I - 1 = I - I = [•- I = 0.25
^ VJ ^
(iii) Expected number of units in the system Ls = — = —-— = — = 1
(X — A, O— o o
(iv) Expected number of units in the queue Lq = - ^ - ^ = j ^ =^ = |
(v) Expected time in minutes customer has to wait in the system
W. = - J — x 60 = 7^™ = ^ = 20 minutes,
(i — A, 6-3 o
.ILLUSTRATIONS 3 —
Customers arrive at a box office window^ being managed by a single individual,
according to a Poisson input process with mean rate of 30 per Hour. The time
requireato serve a customer nas an exponential distribution with a mean of 90
seconds. Find the average waiting time of a customer. Also determine the average
number of customers in the system and average queue length.
[JNTU (Mech., ECE) 1999/CCC)]

Solution :
Arrival process is Poisson i.e., Markovian represented by *M\
Service distribution is Exponential which can be represetned as 'AT (Markovian)
Number of servers = 1
No. of people (customers) allowed to enter queue = °° and queue descipline is
understood as FCFS (First Come First Serve).
Hence, the given queue model is represented as
(M/M/l) : (°°/FCFS)
Queue Models (Waiting Lines) 437

Also given that,

Mean arrival rate = X = 30 per hour
30
= an ^ n per second
60 x 60 r
1
=
"l20F
and service time — = 9 0 seconds

.-. mean service rate (\i) = —- per second

J_
X 20 90 3
Fraction of Busy period or Traffic intensity ( p ) = —= — = —— = — = 0.75
\i 1 120 4
90
Average numbe of customers in the system ( L s )
X p 0/75 0.75 _
= or T J = — = -—7-777 = 7T177 = 3 customers
\i-X 1-p 1-0.75 0.25
Average number of customers in the system ( Ls) = 3 customers
2 2
X o
Average queue length ( L ) = — or - ^ — = 0.75 x 3 = 2.25
H
\L(\i-X) 1-p
Average waiting time of a customer in queue ( W )
= T2- = TTT^r = 4.5 minutes or 270 seconds.

Average waiting time of a customer in the system Ws = — or

X
Wq + service time = 6 minutes or 360 seconds
[-ILLUSTRATION 4
At a certain petrol pump, customers arrive according to apoissonprocess with an
average time of 5 minutes between arrivals. The service time is exponentially
distributed with mean time of! minutes. Find (i) the average queue length (ii) the
average number of customers in the queueing system (Hi) average time spent by a
customer in the petrolpump and (iy) average waiting time of a customer before
receiving the service, fv) if the waiting time in queue is 4 min, a secondpump will
be opened. What should be the arrivatratefor opening secondpump.
[Dr. BRAOU (MBA) 97]

Solution :
This is an (M/M/l) : (oo/FCFS) queue model.
Average inter arrival time = — = 5 min = — hr
A. iA

or X=l2/hr

Average service time = — = 2 min = -—• hr

or \i = 30 / hr
438 Operations Research : Theory and Practice

As X < \i, the steady state solution exists,

Ar 144 4
(i) Average length of queue (Lq ) =^ - ^ = - ^ - ^ = -
A, I9 2
(ii) Average length of system (Ls) = -= -§ = -
\i — K lo o
(iii) Average time spent in the petrol pump (system)
W5 = —^— = - ^ hr = 3.33 min
\i-X 18
(iv) Average waiting time of customer in queue i.e., before receiving the service

(or ^ = ^ - - = 3 . 3 3 - 2 = 1.33 min)

(v) To open second pump W '= 4 min
w ' _ _ 4m m _

X' ±
30 ( 3 0 - A / ) 15
If = 20/hr and inter arrival time=3 min.
ILLUSTRATION 5
A television repairman finds that the time spent on his jobs has a exponential
distribution with mean 30 minutes. If he repairs sets in the order in which they
came in, and if the arrival ofsetsfollow a Poisson distribution approximately with
an average rate of 10 per 8 hour day, what is the repairman's expected idle time
each day? How manyjobs are aheadof the average set just brought in ?
[JNTU (Mech.) 94/P]

ISolulion :
Arrival fashion is Poisson distribution (M); service pattern is exponential
distribution ((i), number of servers is 1, no limit is imposed, repair discipline is FIFO.
Therefore, according to Kendall-Lee, the queue system (M/M/l) : (°°/FIFO)
10 5
and mean arrival rate (X) = — = — sets per hour.

and mean service rate (\i) = — x 60 = 2 sets per hour

H~4 ~8
— < 1, hence steady state can exist.
M-
(i) Expected idle time of repairman each day
number of hours for which the repairer is busy in 8 hour day is
o k o 5 r 1
8 x — = 8 x — = 5 hours
LL o
Idle t i m e for r e p a i r m a n in a n 8 h o u r d a y = 8 - 5 = 3 h o u r s
Queue Models (Waiting Lines) 439

(ii) Expected number of TV sets in the system

5
X o 8 5 „

8
ILLUSTRATION 6
In a railway marshalling yard, goods trains arrive at a rate of 30 trains per day.
Assuming that the inter-arrival time follows an exponential distribution and the
service time (the time taken to dump a train) distribution is also exponential with
an average of 36 minutes, calculate (i) expected queue size (line length ) ,
(improbability that the queue size exceeds 10. if the input of trains increase to an
average of33 per day, what will be the change in (i) ana(ii).
[JNTU (Mech.) 97, OU 83]

Sol ul ion :
From the data, the given queue system has exponential arrival (M), exponential
service pattern (M,) single track (service facility), unlimited queue size and FGFS
discipline.
.*. According to Kendall-Lee, the system is (Af/M/1) : (oo/FCFS)
We have the parameters as
30 1
A. = = — trains per mmute
60 x 24 48 ^
and |i = — trains per minute.

Then traffic intensity (p) = — = — = 0.75 as p < 1 , steady state can exist.
We can proceed with the problem as follows:
(i) Expected queue size (line length)

(ii) Probability that the queue size exceeds 10,

P (n > 10) = p 1 0 = (0.75)10 = 0.056
Now, if input increase to 33 trains per day, then we have
. 33 11 . , 1 . . .
^ = 7 7 7 7 = 777T trams per minute and u. = — trains / min.
60 x 24 480
480 36
Traffic intensity p = ^ = ^ = 0.83
as p < 1 , steady state can exist.
Hence recalculating the values of (i) and (ii), we have

(ii) P(xZ 10) = p l o = (O.83)lo = 0.155 (approx.).

440 Operations Research : Theory and Practice

ILLUSTRATION 7
Arrivals at a telephone booth are considered to be Poisson with an average time of
10 minutes between one, arrival to the next. The length of phone call is assumed to
be distributed exponentially, with mean 3 minutes.
(i) What is the probability that a person arriving at the booth will have to
(h) Whklelephone department will install a second booth when convinced that
arnv would expect waiting for at least 3 minutes for a phone call. By
%n arrival
an
how much
how mi the flow of arrivals should increase in order to justify a second
booth.
(in) What is the average length of queue that forms time to time.
(iv) What is theprobability that it will take him more than 10 minutes altogether to
wait for the phone can and complete his call?
[JNTU (Mech.) 94/CCC, (ECE) 94/CCC]
Solution :
Given problem has Poisson arrival, exponential service time distribution, single
booth (service facility), unlimited queue with FCFS discipline.
T h e r e f o r e in Kendall - Lee notations, it is r e p r e s e n t e d as
(M/M/l) :(~/FCFS)
Also given X = —— persons per minute
, 1
and ^t = — persons per minute

Since, p < 1 the steady state exists, hence we proceed

(i) Probability that a person has to wait at the booth
X

(ii) The installation of second booth can be justified if the arrival rate is more than
the waiting time. Let V be increased arrival rate. Then expected waiting time
in queue will be

A.'= — = 0.16 or ~ = 6 minute.

6 %'
Here Wq = S (given)
1 1 4 1
and increase in the arrival rate, = — - — = — = — = 0.06 arrivals /minutes
b 10 bO 15
(iii) Average length of a busy period
J.
\i 3 0.33 , AQ .
Lb = ~^r = "j 7~ = ~K~v% ~ 1*43 minutes
i~To
Queue Models (Waiting Lines) 441

(iv) Probability of waiting for 10 minutes or more is given by

> 10) = | - (n - X)e~ ^~X)tdt = j (0.3) (0.23) e°'25t dt

10^ 10

J
-0.23(1

10
This suggests that 3% of the arrivals on an average will have to wait for 10
minutes or more before they can use the phone.

Practice Problems
1. In a large computer industry, the average rate of system break down is 10
systems per hour. The idle time cost of a system is estimated to be Rs. 20 per
hour. The working hours per day are 8. The manager of industry considers two
mechanics for repairing. The first mechanic A takes about 5 minutes on an
average to repair a system and demands wages Rs. 10 a hour. The second
mechanic B takes 4 minutes in repairing and charges at rate of Rs. 15 a hour.
Assuming rate of system breakdown is Poisson distributed and repair rate
exponentially distributed which of the two mechanics should be appointed.
Solution :
For mechanic A :
X = 10 Ihr and \L= 12 fhr.
Total cost = total wages + cost of non productive time
= (Hourly rate x number of hours) + (Average number of systems)
x (Cost of idle system / hour) x number of hours.
- 1 0 x 8 + - ^ - x 20 x 8 = 80 + — ^ — x 160 = 880 Rs..
JI-A, 12-10
For mechanic B :

. Total cost = 1 5 x 8 + — ^ - x 2 0 x 8 = 120+ 1Q x 160 = 440 Rs.

\i-X 15 - 10
Though mechanic B charges (440 Rs.) is less, when compared to mechanic^
who charges (880 Rs.) mechanic B can repair three systems more than mechanic^,
i.e., 24 systems for 8 hours (8 x 3). Thus he can save (24 x 20) i.e., 480 Rs. per day, so
that total cost of mechanic B becomes 920 Rs.. (440 + 480).
So, mechanic A has to be appointed.
2. Customers arrive at a one window drive in H.M.T sales counter in a Poisson
fashion with mean 10 per hour. Service time per customer is exponential with
mean 5 minutes. The space in front of the window including that for serviced
car can accommodate a maximum of 3 cars other cars can wait outside.
442 Operations Research : Theory and Practice

(i) What is probability that a customers can drive directly to window.

(ii) What is probability arriving customer have to wait outside indicated space,
(lii) Expected waiting time of customers trucks per day.
Solution :
ryr\

X=\0/hr; \i = -y-= 12/hour.

(ii) P (n > 3) = (X,/jLi)3 = (10/12) 3 = 0.48

(iii) Wq = 5/2 hour or 25 minutes.

3. A scooter mechanic has an average of four customers an hour. The average

service time is six minutes. Arrivals are at Poisson fashion and service is done
at exponential pattern. Find the
(i) Proportion of time during which shop is empty.
(ii) Probability of finding at least 1 customer,
(iii) Average time spent (service time including)
(iv) Average number of customers in system.
Solution :
fin
X = 4 / hour \x = — = 10 / hour; P- x_ 0.4

(i) (ii) ni ]
1 2
(iii) Ws = — hour or 10 minutes
O
(iv) ws 3"
4. The average rate of arrival of trains at a railway station during the busy period
is 20 per hour and actual number of arrivals in any hour follows a Poisson
distribution. The maximum capacity of the railway station is 60 trains per hour
on an average in good weather and 30 trains per hour in bad weather but actual
number of arrivals in any hour follows a Poisson distribution with these
averages. When there is congestion the trains are made to wait for clearance
outside the station.
(i) How many trains would be waiting outside on an average in good weather
and bad weather.
(ii) How long would a train be waiting outside before getting into station in
good and bad weather.
Queue Models (Waiting Lines) 443

Solution:
X = 20; X
| = 60 in good weather and 30 in bad weather.
1 4
(i) L - — in good weather ; — in bad weather

(ii) — hour in good weather ; — hour in bad weather

5. Find the (i) average queue length, (ii) average time spent, (iii) probability that
there would be two customers in the queue in a sales counter manned by one
sales person where the service time is about 6 minute and arrival rate is one
person eveiy 10 minutes.
Solution :
A, = — = 6 / hour, \L = —T= 10 /hour
10 b

(i) Lq = 0.9 customers

(ii) Ws = — hour or 15 minutes

(iii) P2 = (X/\i)2 (1 - X/\i) - 0.144 or 1.44%

6. A maintenance service facility has Poisson arrival rates, negative exponential
service times and operates on a first come first serve queue discipline Break
downs occur on an average of three per day with a range of zero to eight. The
maintenance crew can service on an average six machines per day with a range
from zero to seven. Find the
(i) Utilisation factor of the service facility.
(ii) Mean time in the system
(iii) Mean number in the system in break down.
(iv) Mean waiting time in the system.
(v) Probability of finding two machines in the system.
(vi) Expected number in the queue.
Solution :
X = 3/day; \i = 6/day

s
11 6 3

(iii) Ws - 1 machine (iv) W - — day

9 1

(v) P2 = (k/\ir ( 1 - X/yi) = 0.125 (vi) Lq = — machine

(vii) Idle time = 1 - p = 50%.

444 Operations Research : Theory and Practice

7. Customers for paying electricity bill arrive following Poisson distribution with
an average time of 5 minutes between one arrival and the next. The time of
service is 3 minute and it is at exponential pattern.
(i) What is the probability that service counter is busy.
(ii) How many service points should be established to reduce the waiting
time to less than or equal to half of present waiting time.
Solution :
X = 12/hour; ji = 20/hour
3
(i) Busy period; 1 - Po = X/\i = 0.60 and Wq = — hour

(ii) | ^ ^ ^

Thus number of booths required to achieve new service rate ='^r = 1-4 booths.
8. Customers arrive at a counter in Poisson fashion at an average rate of 15
customers per hour in a single teller counter where bank cheques all cashed.
The service time is exponentially distributed where the cashier takes 3 minutes
to cash the cheque.
(i) Calculate percentage of time the teller is busy.
(ii) Calculate average time customer is expected to wait.
Solution :
X = 15/hour ; p = 60/3 = 20/hr
X 3 e
(i) Busy period = 1 -P0~~~~A i* -> teller *s ^ u s y f° r 7b% of time.

(ii) Ws = — hour or 6 minutes.

9. On an average 96 patients per 24 hour day require the service of an emergency
clinic. Also on average, a patient requires 10 minutes of active attention. Assume
facility can handle only one emergency at a time. Suppose it costs the clinic Rs.
100 per patient treated to obtain an average service time of 10 minute and each
minute of decrease in this average time would cost Rs. 10 per patient treated.
How much would have to be budgeted by the clinic to decrease average size of
queue from 1 V3 patients to V2 patient.
Solution ;
X = 90/day ; p = ( 2 4 x 6 0 > = 144/day

4
(1) Lq = - patients;
4 1
When L is changed from -r to —
00

.„, , 1 }? 96 2
Then p changes to —• = —-—-——- = ——;——-
2 (i (p, - X) \i (fi - 96)
p,' = 192 patients a day.
Queue Models (Waiting Lines) 445

New service time - 24 x 60/192 = 7.5 min.

Reduced time = 2.5 min
Increased cost due to decreased service time = 2.5 x 10 = Rs. 25
Total cost = 192 x 100+25 = Rs. 19225
Present cost = Rs, 144 x 100 = 14400
Increased budget = Rs. 4825
10. A cashier at a bank works on exponential fashion. At what average rate must
he work in order to ensure a probability of 0.9 so that a customer will not wait
longer than 12.minutes. There is only one counter where customers are served
at rate of 15 per hour who are arriving at Poisson fashion. [JNTU (CSE) 2001]
Solution :

6 Q -4 . v ./>12) = 0,9 =* P(W> 12) = O.\

P [(waiting > 12]= J - f\i - V\e~ ^~X)txdt = 0.10(given)

Thus es ~~12 ^ = 0.4 (i or — - 2.48 min/service.

11. Oil tankers arrive at a refinery to transport oil for distant markets. They are
served on first come first serve fashion. Tankers arrive at a rate of 10 per hour
whereas loading rate is 15 per hour. Arrivals are in Poisson fashion filling in
exponential fashion. Transporters are complaining that their trucks have to
wait for nearly 12 min at plant. Examine whether complaint is justified. Also
Jetermine probability that loaders are idle in above problem.
Solution :
%= 10/hours; \i = 15/hour
(i) W = — hour or 8min, hence complaint is not jsutified.

(ii) Idle time Po = 1 - p = —- or 33.33%.

12. Consider a self service store with one cashier. Assume Poisson arrivals and
exponential service times suppose that nine customers arrive at an average
every 5 minutes and cashier can serve 10 in 5 minutes. Find
(i) Average number of customers in queuing for service.
(ii) The probability of having more than 10 customers in system,
(iii) Probability that a customer has to want in queue for more than 2 minutes.
If the service can speed up to 12 in 5 minute by using different cash register,
what will be effect on quantities (i), (ii) and (iii).
446 Operations Research : Theory and Practice

Solution :
Case I :
9 .12.
5' 5
(i) Ls = 9 customers; (ii) P(n> AIO

oo

l X)t
(iii) P (waiting ~ A =0.67.
9 t*1

Case II:
- 9 12

(i) L5 = 3 customers. (ii) p (n > 10) = (0.75)10

oo

iii) P (waiting > 2) = J - (ja - X) e~ * " x ^ = 0.30.

2 H

13. In a tool crib manned by a single assistant, operators arrive at the tool crib at
a rate of 10 per hour. Each operator needs 3 minutes on an average to be served.
Find out loss of production due to time lost in waiting for an operator in a shift
of 8 hours if the rate of production is 100 units per shift.
Solution :
X= 10/hour; |i=20/hour
1 8
(i) ^q~~^R hour; average waiting time per shift = — hour.
c , • , -. 2 100 c .
Loss of production due to waiting = — x — - = 5 units.
u o
14. A fertilizer company distributes its products by trucks loaded at its only loading
station, Both company and contractor trucks are used for purpose. It was found
that on an average every 5 minutes one truck arrived and average loading time
was 3 minutes 40% of trucks belong to contractors. Make probable assumptions
and find out
(i) Probability that a truck has to wait,
(ii) Waiting time of truck that waits,
(iii) Expected waiting time of customers truck per day.
Solution :
* = ^ = 12/hour; JLL = ^ = 20/hour.
5 3

(i)
Queue Models (Waiting Lines) 447

(ii) Ws = — hour or 7.5 minute.

o
(iii) Total waiting time = number of trucks per day x (% contractors truck)
x expected waiting time for truck.
40 A
= 12 x 24 x - — x =8.64 hours/day.
100 JLX (JLX — A-)
15. Customers arrive at a first class ticket counter of a theatre in a poisson
distributed arrival rate of 25 per hour. Service time is constant at 2 minutes.
Calculate
(i) Mean number in waiting time,
(ii) Mean waiting line.
Answer: (i) 5 people (ii) 10 min.
16. At a public telephone booth in post office arrivals are considered to be Poisson
with an average interval time of 12 minutes length of call is distributed
exponentially with average time being 4 minutes. Calculate
(i) Probability that fresh arrival need not wait.
(ii) Arrival have to wait more than 10 minutes before he gets chance,
(iii) Average length of queue that form from time to time.
Answer: (i) 0.67 (ii) 0.063 (iii) 1.5
17. Weavers in a textile mill arrive at a department store room to obtain spare parts
needed for keeping the looms running. The store is manned by a single
attendant. The average arrival rate of weavers per hour is 10 and service rate
per hour is 12. Both arrival and service rate follow Poisson process.
Determine :
(a) Average length of waiting time.
(b) Average time a machine spends in the system.
(c) Percentage of idle time of department store room (attendant).
[JNTU (CSE) 99]

Answer: (a) 25 min (b).30 min (c) 40%

18. In the production shop of a company the breakdown of the machines is found
to be poisson with an average rate of Rs. 3 machines per hour. Break down time
at one machine costs Rs. 40 per hour to the company. There are two choices
before the company for hiring the repairman. One of the repairman is slow but
cheap, the other fast but expensive. The slow-cheap repairman demands Rs.
20 per hour and will repair the broken down machines exponentially at the
rate of 4 per hour. The fast-expensive repairman demands Rs. 30 per hour and
will repair machines exponentially at an average rate of 6 per hour, which
repairman should be hired? [JNTU Mech/Prod/Chem/Mechatronics 2001/S]
448 Operations Research : Theory and Practice

Model-2: (M/M/l) : (°°/SIRO)

Single Channel - Unrestricted Queue: This model is identical with model - 1 in
all respects except for the queue discipline. Since the operating characteristics of a
queue model are independent of queue discipline (as the derivation for probability is
independent of queue discipline), all the results remain unchanged. Therefore the
problems under this model may be solved using same formulae as those of model -1.

8.11 Model-3: (M/M/l) : (N/FCFS)

Single Channel- Restricted Queue Model: This model differs with model-1 in
respect of limit of queue. Here the capacity of the system is limited to TV. The relevant
formulae are given below :
P
I pnpn forO<n<N,

for p = 1 or X = \i
N+
and Po = ——- ; p * 1 and p < 1
1- p
Where p=—

N = limit of queue system.

n = number of customers arrived.
2. Expected size of system or expected number of customers in the system

T - P

= — for p=1

3. Expected length (size) of queue or expected number of customers waiting in

queue.

L =L
i *-\

4. Expected waiting time of customer in the system \Wt + —

5. Expected waiting time in queue

Queue Models (Waiting Lines) 449

ILLUSTRATION 8
Dr. Raju rs out-patient clinic can accommodate six people only in the waiting hall.
The patients who arrive when hall is full, balk away. The patients arrive in Poisson
fashion at an average rate of 3 per hour and spend an average of15 minutes in
Doctor rs chamber which is exponentially distributed. Find
(i) Theprobability that a patient can get directly into the doctor fs chamber upon
his arrival
(ii) Expected number of patients waiting for treatment.
(Hi) Effective arrival rate.
(iv) The time a patient can expect to spend in the clinic.
Solution :
From the above problem,
Arrival: Poisson (M); service : Exponential (Af);
Number of servers : One; (1) discipline : FCFS;
But the queue is limited
The model is (M/M/l) : (N/FCFS)
Now capacity of the system
TV = hall capacity + doctor's chamber capacity
= 6+ 1 = 7
Arrival rate (k) - 3 per hour.

Service rate (n) = —- per min = 4 per hours.

X 3
p = — =4— = 0.75 i.e., < 1 (stead state exists)
M-
(i) Patient will directly enter doctor's chamber if the hall is empty on his arrival.
The probability at this situation is Po

-=0.2778

(ii) Expected number of patients waiting for treatment is (Lq)

8 -,

= 2.1

(iii) Effective arrival rate (Xeff) = \i (1 - Po)

= 4 ( 1 - 0.2778) = 2.89 = 3 per hour.
450 Operations Research : Theory and Practice

(iv) The time the patient can expect to spend in the system
Lq + (l-Po) 1.36+ 1-0.2778
= 0.72 hours
2.89
- 43.2 minutes.

Practice Problems
1. Consider a single server queuing system with Poisson input, exponential service
times. Suppose the mean arrival rate is 3 calling units per hour, the expected
service time is 0.25 hour and the maximum permissible number calling units
in the system is two. Derive the steady-state probability distribution of the
number of calling units in the system and then calcualte the expected number
in the system.
Answer : The expected number of calling units in the system = 0.81.
2. If for a period of 2 hours in the day (8 to 10 a.m.) trains arrive at the yard at
every 20 minutes but the service time continues to remain 36 minutes, then
calculate for this period.
(a) The probability that the yard is empty, and
(b) The average number of trains in the system, on the assumption that the
line capcaity of the yard is limited to a four trains only.
Answer:
X = 1/20; ji = 1/36; p = 36/20 = 1.8 (> l ) a n d n = 4
p-1
(a) = 0.04
N+\

(b) Ls = = 2.9^3.
71=0

8.12 Model - 4 (M/M/S) : (-/FIFO)

Multi Channel - Unrestricted Queue Model: This model can be supposed as the
extension to model-1. Here instead of single channel, multi channel (multiple servers)
in parallel say V lines are assumed. With other notations as usual we use the following
formulae for this model.
p
1. P - P for 1 <n<s

for n>s

Where p=—
n
5-1 ~s~\
-1 ^ I (X\ 1 (X 5(1
\n Js 1 - p or ' s|i- X
n=0
Queue Models (Waiting Lines) 451

2. Expected length of queue

1 (X) X\i
lfr-1) '[MJ \s\i~X)
3. Expected number of customers in the system

4. Expected waiting time of the customer in queue

'*" X r-1) hi
5. Expected waiting time in' the system

— or
6. The probability that a customer has to wait (busy period)

P(n>s) =~
1
ILLUSTRATION 9
A super market has two girls ringing up sales at the counters. If the service time
for each customer is exponential with mean 4 minutes, and if people arrive in a
Poissonfashion at the rate oflO/houn
(a) What is the probability of having to wait for the service.
(b) What is the expected percentage of idle timefor each girl?
(c) Find the average length and the average number of units in the system ?
[JNTU CSE 97/S JNTU - Mech/Prod/Chem/Mechatronics 2001/S]

Solution :
No. of servers (s) = 2, Arival rate ( X) = — and service rate ((i) = — per minute],

1
S \L

Therefore,

2-1

2! 6

1 1
452 Operations Research : Theory and Practice

(a)

=y p = y
n=2 n=2'

= ~ (or) 0.167
X 1 2
(b) The expected idle time for each girl is 1 -"— i.e., 1 - — (or) — ( = 0.67).
Hence, the expected percentage of idle time for each girl is 67%.
(c) Expected length of customers waiting time = —
1
or 3 minutes.
0.50-0.17
I L L U S T R A T I O N 10
A bank has two tellers working on savings accounts. The first teller handles
withdrawals only. The second teller handles deposits only. It has been found that
the service time distribution for deposits and withdrawals both are exponential
with mean service time 3 minutes per customer. Depositors arefound to arrive in
a Poisson fashion through out the day with mean arrival rate 16 per hour.
Withdrawals also arrive in a Poisson fashion with mean arrival rate 14 per hour.
What would be the effect on the average waiting time for depositors and with
drawers, if each teller could handle both withdrawals and deposits? What would
be the effect of this could only be accomplished by increasing the service time to
3.5 minutes?
oolulion :
Initially, we can assume two independent queueing systems as follows.
Withdrawers System Depositors System
Arrivals : Poisson (M) Poisson (Af)
Service : Eponential (M) Exponential (M)
Number observers: One (1) One (1)
Limit of queue : Unlimited (°°) Unlimited («>)
Discipline : FCFS FCFS
Model : (M/M/l) : ( ~ / FCFS ) (M/M/l) : ( «, / FCFS)
Mean arrival rate : (XA= 14/hours X9 = 16/hours

Mean service rate : = — per min 20/hours j^2 = T p e r m m 20/hour

o

service time — = 3 min ) (Mean service time — =3 min)

Average waiting time Average waiting time

14 16
W =
q 2 0 ( 2 0 - 14) 20(20-16)
= X=7 min = -rhrs= 12 min
5
Queue Models (Waiting Lines) 453

Now Consider Combined System :

Arrivals : Poisson (M)
Service : Exponential (M)
No. of servers : two (2)
Limit of queue : Unlimited ( °°)
Discipline : FCFS
.-. Model : (M/M/2) : ( « / FCFS)
Mean arrival rate : ^ = 1 4 + 1 6 = 30/hours
Mean service rate : [I - 20/hours
X 30 3
Number of servers (5) = 2 and p = — = -—— = —
s LI 2 x 20 4

X 30 __3
Also
H ~ 20 " 2
n s -1
Vi
Now
n=0 V

2 n-1

V — — — — r 40
n! 2 2! 2 40-30

1
T1
1 9 40
+
2xlX4X10j

Average waiting time of arrivals in the queue

1 m _y_

1 20
(2 - 1)! • [ 2 j ' (40 - 30) 2 ' 7
9 20 1 9 ,
=
4xT5o7 =T^hours-
= 3.86 min
454 Operations Research : Theory and Practice

Combined waiting time with increased service time when

%' - 30 /hour and — =3.5 min
60 120
or = — = — per hour
30 30
= -i.e.,<l and —=
S\L' 2(120/7) 8 \L' 120/7 4
-,-1

I (1) (1 2 x (120/7)
i\\ [I 4a j I 29i! [I 4Aj {2 (120/7)-30}
O l1 -pi
1 49 2(120/7)
XX
0! 9 +
1! 4 +
2 x l 1 6 XX (30/7)

Average waiting time of arrivals in the queue

W
* (s-iy.
1 120/7
(2-1)1 ^4j [2(120/7)-3Q] 2 15
49 120/7
(30/7) 2 X
16XX 7
1 5 " 16 X
30 X
15
343
hours
30x60
343
min =11.433 min.
30

Practice Problems
" 1. A post office has 3 windows providing the same services. It recieves on an
average 30 customers/hours. Arrivals are Poisson distributed and service time
exponentially. The post office serves on average 12 customers/hours.
(i) What is the probability that a customers will be served immediately?
(ii) What is the probability that a customer will have to wait?
(iii) What is the average no. of customers in the system.
(iv) What is the average total time that customer must spend in the post
office?
Answer :(i) 0,29 (ii) 0.70 (iii) 6 customers (approx) (iv) 12 min.
Queue Models (Waiting Lines) 455

2. Two repairmen are attending five machines in a workshop. Each machine

breaks down according to a poisson distribution with a mean 3 per hour. The
repair time per machine is exponential with mean 15 min.
(i) Find the probability that the two repairmen are idle and that one
repairman is idle.
(ii) What is the expected number of idle machines not being served?
3. Given an average arrival rate of 20 per hour; is it better for a customer to get
on service, at a single channel with mean service that of 22 customers or at one
of the two channels in paralle with mean, service rate of 11 customers for each
of the two channels? Assume that both the queues are of M/M/S type.
[JNTU (Mech.) 2001]

Answer : Better to get service at two channels.

5. A telephone exchange has two long distance operators. The telephone
company finds that during the peak load, long distance calls arrive in a Poisson
fashion at an average rate of 15 per hour. The length of service on these calls
is approximately exponentially distributed with mean length 5 min.
(a) What is the probability that a subscriber will have to wait for his long
distance call during the peak hours of the day?
(b) If subscribers will wait and are served in turn, what is the expected waiting
time?
Answer : (a) 0.48 (b) 32 min.
6. A company currently has two tool cribs, each having a single clerk, in its
manufacturing area. One tool crib handles only the tools for the heavy
machinery. While the second one handles all other tools. It is observed that for
each tool crib the arrivals follow a Poisson distribution with a mean of 20 per
hour and the service time distribution is negative exponential with a mean of
2 minutes. The tool manager feels that, if tool cribs are combined in such a way
that either clerk can handle any kind of tool as demand arises, would be more
efficient and the waiting problem could be reduced to some extent. It is believed
that the mean arrival rate at two tool cribs will be 40 per hour, while the servive
time will remain unchanged.
Compare in status of queue and the proposal with respect to the total expected
number of machines at the tool crib(s), the expected waiting time including
service time for each machine and probability that he has to wait for more than
five minutes.
Answer: For separate tool cribs: 2 arrivals, 6 min.
For combined tool cribs : 2.4 arrivals, 3.6 min.
456 Operations Research : Theory and Practice

7. In a machine maintenance a mechanic repairs four machines. The mean time

between service requirement is 5 hours for each machine and forms an
exponential distributionvThe mean repair machine down time costs Rs. 25 per
hour and the machine costs Rs. 55 per day of 8 hours.
(i) Find the expected number of operating machines,
(ii) Determine the expected down time cost per day.
(iii) Would it be economical to engage two machines each repairing only two
machines?
8. A petrol pump station has two pumps the service times follow the exponential
distributions with a mean of 4 min and cars arrive in a Poisson distribution at
a rate of 10 cars per hour. Find the probability that a car has to wait. What
proportion of time the pump remain idle?
Answer: 0.167; 67%
9. A two-channel waiting line with Poisson arrival has a mean arrival rate of 50
per hour and exponential service with a mean service rate of 75 per hour for
each channel. Find
(a) the probability of an empty system
(b) the probability that an arrival in a system will have to wait.
Answer: 0.83,0.167
10. A telephone company is planning to install telephone booths in a new air port.
It has established the policy that a person should not have to wait more than
10% of the times he tries to use a phone. The demand for use is estimated to
be poisson with an average 30 per hour. The average phone call has an
exponential distribution with a mean time of 5 min. How many phone booths
should be installed.
Answer : 6 phone booths

Review Questions
1. Write short notes on waiting line situations. [CSE - M 97]

2. What are waiting line costs ? Explain them with the aid of graph.
[Mech - 98/S, MBA - BRAOU - 98]

3. Write short note on queuing models [Mech 98/S, MBA - OU - 98]

4. Write short note on basic structure of queuing model[Mech. 99 M, Mech - 2000 M]

5. Discuss queuing system [Mech 99/S, MBA - IGN0U - 97]

Queue Models (Waiting Lines) 457

6. Briefly explain the important characteristics of a queuing syster(Mech 2000 - OR]

7. Briefly explain the representation of queue models based on Kendall-Lee

notations. [Mech - 97/P]

8. Discuss the operating characteristics of basic queue model [CSE - 96/C]

9. Discuss different queue disciplines used in waiting lines models. [Mech - 98/C]

10. Discuss the arrival and service patterns used in Q-models MBA - 0U - 98]

11. Explain various queue configurations [MBA - IGNOU - 96]

12. What do you understand by (M/M/l) : (~/FCFS). Explain the terms

[MBA IGNOU - 97, 99, 2001]

13. Discuss the customer behaviours that are often found in queue systems.
[MBA - IGNOU - 98]

14. Discuss important features of a waiting line model [Mech - 96/P, MBA - 0U - 98/S]

15. Explain, the applications of queue models in business situations.

[Mech - 96/C, MBA - IGNOU 97, BRA0U - 95]

Objective Type Questions

L A customer buying tickets in a black market at a cinema hall is said to be
(a) balker (b) reneger
(c) jockeyer (d) dissatisfied
2. An expediting in production shop in example for queue discipline
(a) FIFO (b) LIFO
(c) SIRO (d) pre-emptive
3. The dead bodies coming to a grave yard is an example of
(a) pure death process (b) pure birth process
(c) birth and death process (d) not a queue process
4. In (M/M/S) : (N/FIFO), which of the following is wrongly stated
(a) poisson arrival (b) exponential service
(c) single server (d) limited service
5. Which of the following is not considered as the negative behaviour of customer
according to queue disciplines
(a) reneging (b) jockeying
(c) balking (d) boarding
458 Operations Research : Theory and Practice

6. A steady state can exist in a queue if

(a) X > |x (b) k< (i
(c)X<v (d)XZ)L
7. The system of loading and unloading of goods usually follows
(a) FIFO (b) LIFO
(c) SIRO (d) pre-emptive
8. If the operating characteristics of a queue are dependent on time, then it is said
to be
(a) steady state (b) transient state
(c) busy period (d) explosive state
9. A person who goes out of queue by losing patience to wait is said to be
(a) reneging (b) jockeying
(c) balking (d) boarding
10. An office filing system follows
(a) FIFO (b) LIFO
(c) SIRO (d) non-pre-emptive
11. The queue discipline in stack of plates is
(a) FIFO (b) LIFO
(c) SIRO (d) non-pre-emptive
12. The queue discipline in selecting the contestants for a game such as "Kaun
Banega Corodpathi (KBC)" is
(a) FIFO (b) LIFO
(c) SIRO (d) pre-emptive
13. SIRO discipline is found generally in
(a) office filing (b) trains arriving to platform
(c) lottery (d) loading and unloading
14. A Poisson arrival, exponential service by single server to limited queue selected
randomly is represented as
(a) (M/E/S) : (-/SIRO) (b) (M/M/S) : (N/SIRO)
(c) (M/M/l) : (N/SIRO) (d) (M/M/l) : (~/SIRO)

15. For a simple queue (M/M/l), p =.— is called

(a) traffic intensity (b) fraction of busy period
(c) utilization factor (d) any of the above
Queue Models (Waiting Lines) 459

16. The unit of traffic intensity is

(a) Poisson (b.) Markov
(c) Erlang (d) Kendall
17. In a (M/M/l) : (oo/FCFS) model, the length of the system (Ls) is given by
2 ^2
(a) ~r~~ ( b ):—TT

(d)-

18. In (M/M/l) : (oo/FIFO) model, represents

(\i - A)
(a) length of system (Ls) (b) length of queue (Lq)
(c) waiting time in queue (W ) (d) waiting time in system (Ws)

19. Which of the following model is same as that of (M/M/l) : (oo/FIFO)

(a) (M/M/l) : (N/FIFO) (b) (M/M/S) : (~/FIFO)
(c) (M/M/l) : (oo/SIRO) (d) none
20. The characteristics of queue model are independent of
(a) service pattern (b) number of service points
(c) limit of queue (d) queue discipline
21. Which of the following is wrong relation in a simple queue model

22. By seeing the large queue size, a customer may pass out without entering the
queue thinking that he may not get service. This is called
(a) walk - out (b) balking
(c) jockeying (d) reneging
23. A doctor rushing to an emergency case leaving his regular service is said to be
(a) pre-emptive£) discipline (b) non-pre-emptive discipline
(c) reneging (d) balking
24. A service system in which customer is stationary but server is moving is found
with
(a) buffet meals
(b) person attending breakdown of heavy machines
(c) vehicles at petrol bunk
(d) out-patients at doctor's clinic
460 Operations Research : Theory and Practice

25. The waiting time in system of a simple queue model is equal to

26. Mean arrival rate at a telephone booth is Poissonal with 6 per hr and average
phone call time is distributed exponentially at 3 min. Then the probability that
a person arriving has to wait is

(a) | (b)0.3

(c) 0.6 (d) 0.1

27. On an average 30 goods trains arrive at a station Poissonally loading time is
distributed exponentially at 36 min. Then average number of trains in queue
will be
(a) 2 (b) 3
(c) 4 (d) none of the above
28. A T.V repairer spends about 30 min on each job exponentially distributed, and
is busy for 5 hrs in a 8 hr day. The mean arrival rate is
5
(a) 10 per 8 hr day (b) - per hr

(c) 10 per hour (d) 6 per hour

29. In a Four teller bank counter, service rate is exponential with 3 per hr while
customers arrive at Poisson rate of 10 per hr. Then utilization factor is

ccf <
30. Given Lq = 3.24customers X = 10/hr ; ^i = 3/hr ; Ws =
(a) 32.4 min (b) 35.4 min
(c) 32.73 min (d) 39.42 min

Fill in the Blanks

1. Customers losing patience out of inordinate delay in service walk out of queue
system, are said to be
2. Pre paid taxi at an airport follow queue discipline
3. The steady state can exist in queue system if X |i
Queue Models (Waiting Lines) 461

i. The storage or memory of files in a computer (RAM) in accordance with

type of queue discipline
5. The unit of utilization factor is
6. The state of queue system whose characteristics are time dependent is
state
7. A Poisson arrival, exponential service, 2 server system with infinite input at first
come first served is represented in Kendall-Lee notations as
8. In a 20-bed hospital, a doctor makes rounds and treats in the exponential
distribution according to the severity of the desease. If impatients are expected
in Poisson fashion, the queue model is represented in Kendall -Lee notations
as
9. A medical representative when went to an out-patient clinic of a doctors, he was
asked to wait till the current service in the doctor's cabin is finished, and then
was given appointment. This is type of queue discipline.
10. The dresses and cloths kept in a suitcase follow queue
discipline
11. For [(M/M/1): (WFIFO)] model, the length of queue (L ) in terms of busy period
(P) is
12. The characteristics of queue model are independent of

13. Value of ^—^ of [(M/M/l) : (-/FIFO)] model states of

(1 - P)2
queue length
14. The cost of waiting per unit time with increase of service
level.
15. Given traffic intensity = 0.75, the server will be idle for hours in
a 8 - hr day
16. A lottery system follows queue discipline
17. For (M/M/C) : («VFIFO) model, with the parameters as arrival rate (k) and
service rate (|i), the traffic intensity (p) =
18. An Erlangian service distribution for Poisson arrival of single server in a limited
size queue selected at random is represented in Kendall-Lee notation as
19. The cost of operating a service factility per unit time with the
increase of level of service
20. Aeroplanes landing on a run way at an aerodrom follow queue
discipline
462 Operations Research : Theory and Practice

Answers
Objective Type Questions :
l.(c) 2(d) 3. [b) 4.(c) 5. (d)
6.(c) 7.(b) 8. (b) 9. (a) 10. (b)
11. (b) 12. (c) 13. (c) 14. (c) 15. (d)
16. (c) 17. (c) 18. (c) 19. (c) 20. (d)
21. (b) 22. (b) 23. (a) 24. (b) 25. (b)
26. (b) 27.(b) 28: (a) 29. (d) 30. (d)
Fill in the Blanks :
1. reneging 2FGFSorFIFO
3. < (less than) 4. SIRO
5. Erlang 6. transient
7. (M/M/2) : (~/FCFS) 8. (M/M/l) : (N/priority)
9. non-per-emptive 10. LIFO
12. queue discipline
1L
w
13. variance or fluctuation 14. decreases
15.2 16. SIRO
18. (M/Ek/1): (N/SIRO)
17.—

19. increases 20. FIFO or FCFS

Sequencing 465

9.0 Introduction
It is common problem in the production process to find the sequence of the
jobs that will result in least idle time for the better utilisation of equipment. However,
it is not applicable for those operations where the technological sequence is already
fixed.
For example, one has to wear socks before wearing shoe. Here, the sequence
is fixed where as the sequence of jobs in which they have to be processed (the order
of jobs is not fixed) is important when some jobs are awaiting the operations with
different timings (the technological sequence of operation is assumed to be fixed).
For example there are five books to be printed and bound. Here, the operational
sequence i.e., printing and then binding is fixed but the sequence in which the books
are to be taken for operations is not fixed. The right sequence depending on the time
required for each operation will minimise the idle time of printer and binder. This
chapter deals with such techniques to find the right sequence that minimises the idle
time of operations.
It will be most convenient to study the job sequencing in the following models.
1. ^jobsX 1 machine (or man or department)
2. n jobs X 2 machines
3. n jobs X 3 machines
4. njobsXm machines
5. 2 jobs X n machines
These are explained here below.

9,1 Sequencing ofn Jobs X1 Machine

When different letters are to be typed by a typist, he has to decide the order in
which he has to type. When a student has some questions to answer in an exam, he
would first choose the order in which he has to present to make it effective. When
certain jobs are to be done in a production shop on one machine only, then it is the
turn of the engineer to decide the order of these jobs to process. If only one operator
is there to paint some cars in a automobile repairing shop, he decides the sequence
of the cars to be painted. In all such situations one has to follow certain discipline or
rule. The following are the rules from which a person can select for his job sequencing
when only machine is available. (The term 'machine' here is used in wide sense to
mean anything like men, material, department, equipment or any relevant one as the
case maybe).
1. First In First Out (FIFO) or First Come First Serve (FCFS) : When the jobs
do not require any preferential treatment, this rule is considered. Complaints
at telecom department, electricity departments, trains on single track or
platform, customers at a retailer or a telephone booth etc., will follow this rule.
466 Operations Research : Theory and Practice
2. Last In First Out (L1FO) : When a machine is dismantled for repair or
overhaul, the parts are put back in LIFO system. Pipeline laying in water works,
electrical wiring maintenance system will follow this model. The office filling
system, loading and unloading of trucks, dressing / undressing shirt and a coat,
wearing/removing socks and shoe are some more common examples that occur
in daily life to follow LIFO.

3. Earliest Due Date (EDD): Most of the times the production departments are
asked for the probable time of completion of the job and based on these
promises the production engineer plans his shop production.

4. Shortest Process Time (SPT) : This is another policy to apply on the

engineering jobs with the concept that the jobs those take less time to perform
are taken first. A student while writing his exam would prefer to write the
question that takes lesser time to that takes longer duration.

5. Longest Process Time (LPT) : In contrast to the above, the job takes longer
time will be taken up first and the jobs which take smaller time will be taken up
last.
6. Pre-emptive Priority Rule : When a job is very urgent, it will be taken up on
priority basis by attending immediately stopping all other jobs. Under this rule
the highest priority job is allowed to enter into the service immediately even if
another job with lower priority is already in service. Doctors use this discipline
when emergency cases arrive to their clinics.

7. Non Pre-emptive Priority Rule : In this case, highest priority goes ahead in
sequence but service is started immediately after completion of the current
service. For example, a doctor gives priority to a medical representative, but
he gives the appointment after the finishing the current job, but will not stop
the current service unlike in an emergency case.

8. Priority Service : Priority is given to certain jobs by the virtue of their

importance or recommendation of top officials or expediting jobs etc.
9. Select In Random Order (SIRO) : Under this rule the jobs are selected at
random for operation irrespective their arrival, urgency, due date etc.
10. Minimum Cost Rule : The jobs are selected in the ascending order of their costs.
11. Maximum Profit Rule : The jobs are selected in the descending order of the
profits.
In the above the first five are static in nature and the rest are dynamic

92 Sequencing of n Jobs X2 Machines: Johnson's Rule

S.M.Johnson suggested a sequencing rule for a situation where there is a group
of n jobs to be processed through two successive work centers. The rule ensures
minimum completion time for the group of n jobs by minimising the total idle times
of the work centers.
Sequencing 467

9*2A Assumptions for n Jobs X2 Machines Sequencing

The following assumptions are usually made while dealing with sequencing
problems.
1. No passing rule is observed strictly i.e., the same order of jobs is maintained
over each machine. In otherwords, a job can not be processed on 2nd machine
unless it is processed on 1st machine.
2. Only one operation is carriedout on a machine at a time.
3. Processing times are known and deterministic.
4. Processing time of a job on a machine is independent of other jobs.
5. The time involved in moving jobs from one machine to another is negligible
6. Each operation, once started, must be completed.
7. An operation must be completed before its succeding operation can start.
8. One machine of each type is available.
9. Ajob is processed as soon as possible, but only in the order specified.
10. All the men/machines work with consistent efficiency
11. Setting time is either included in process time or neglected
12. No reworking is allowed.

9.22 The Rule

The rule is
1. Identify the job with lowest processing time among all the jobs (2n) on both the machines.
2. If this shortest processing time belongs to the first machine (or work centre), then this job
is placedfirst in sequence, if the shortest processing time belongs to the second work centre,
this job is put last in the sequence.
3. For the remaining jobs, repeat step 2y until all the jobs are placed in the sequence.
Following example illustrates Johnson s rule.
ILLUSTRATION 1 ~
Venkatesh automobile works has six cars for repair. The repair consists of two
step procedure viz* Dent removing and pointing. The time estimates are as
follows:
CarNumber 1 2 3 4 5 6
Time estimate (dent removing) 16 10 11 13 8 18
Time estimate (painting) 15 9 15 11 12 14
Find the correct sequence of operations and prepare Gantt chart and time
schedule chart so as to minimise the idle time. [JNTU (Mech.) 92/P]

Sequencing :
Step 1 : The most minimum time among all 12 jobs (6 dent removing jobs and 6
painting jobs), is taken by car number 5 (say C5) for dent removing i.e., 8 hours.
468 Operations Research : Theory and Practice

Step 2 : As the least time 8 for C5 falls under first operation, this is to be sequenced at
extreme left.

Now this C5 is deleted for further sequencing.

Now the problem is
Cars 1 2 3 4 S^$:^-3 6
Dent. 16 10 11 13 | g j f
:
18
Paint 15 9 15 11 ^/j$^ 14

Step 3 : Step - 2 is to be repeated till the sequencing is completed.

As the next shortest processing time i.e., 9 for C2 belongs to the second work
centre (i.e., painting) this is to be sequenced to the extreme right.

Cars 1 -.•"."" 9 - V - ; 3 4 6
Dent. 16 11 13 18
Paint 15 15 11 14
(The next minimum is 10 for, again C2, which is already sequenced and deleted
from further consideration). Then next minimum is 11, which appears in C3
for work centre - 1 as well as in C4 for work centre 2. Therefore C3 goes to left
and C4 goes to right available positions.
c2

Cars 1 6
Dent. 16 18
Paint 15 14

(Similarly, the next shortest processing times are 12 for C5 for work centre-2
and 13 for C4 for work centre-1, but these are already allocated and deleted
for the consideration). Thus C6 occupies next position towards right side of the
left out positions as this has got the next least process time i.e., 14 hours at work
centre-2.
c4 c2

c^ars
16
]
Mi
Dent.
iiii
Paint 1
5 alii
Sequencing 469
O b v i o u s l y , C j c o m e s i n t h e left o u t p o s i t i o n . T h u s t h e s e q u e n c e is . . .

c4 c2
Schedule Chart:
Car First Work Centre Second Work Centre Idle Remarks
No. (dent removing) (Painting) Time
Time In Job Time Time Out Time in Job Time Time Out
c5 0 8 8 8 12 20 8* Painter can start at 8th
hr. only
c3 8 11 19 20 15 35 0 C3 waits for 1 hour.
c, 19 16 35 35 15 50 0 No waiting for job and
painter
35 18 53 53 14 67 3 Painter waits for 3 hours.

c4 53 13 66 67 11 78 0 C4 waits for one hour.

c2 66 10 76** 78 9 87 0 C2 waits of 2 hours.
76 87 Total Painter may be asked to
Total elapsed
idle come 8* hours late to the
time
time duty, and denting
= 11 removing operator 11
hrs early.
* Some authors treat this as slack time but not the idle time, as painter may be
asked to come late to the duty 8 hours hence to the dentor. If this duration is
not considered as the idle time of waiting time for painter, the idle time is only
3 hours.
** Similarly the first operator, i.e., dentor can be asked to go early and need not
stay till 87 hours and hence the difference 11 hours is not treated as idle time.
Thus the delivery schedule may given as Cj on 50th hr, C2 on 87th hr, C3 on
35th hr, C4 on 78th hr, C5 on 20th hr and C6 on 67th hr so as to minimise the idle
times for the machines or the repair men.
Gantt Chart:
* Idle time .
|0 / 8 20 35 50\53 67 78
Painting IQ,
(12) (15) (15) (14) (15) (9)
0 8 19 35 53 66 76
Denting

(8) (11) (16) (18) (13) (10)

1 1 1 1 1 1
10 20 30 40 50 60 70 80 90
FIGURE?.! : GANTT CHART
470 Operations Research : Theory and Practice

Remarks:
1. If there is same minimum processing time for two jobs, one at work centre-1
and other at work centre-2 then, the one with shortest time at work centre-1 is
kept at most left and the other at most right.
2. If there is a tie in shortest processing time for any two jobs in first work centre,
then the sequence is to be determined by checking their minimum time at the
second work centre. The job having minimum time is to be allocated most right
and the other is kept just before it.
3. If the tie is for two jobs at second work centre, check their minimum time at
first work centre and the job having minimum time is to be given left position
and the other just after it.
4. If there is a tie in minimum processing times of two jobs with both the machines
or work centers then the sequencing is done by selecting arbitrarily. In this case
the problem will lead to multiple or alternate optimal solutions (more than one
optimal solution).
5. If there is a tie i.e., with same minimum process time for a particular job in
both the series of machines, then also, we get an alternate optimal solution, as
it can take either left or right available position.
6. If the profits are given, then it is maximisation case for which we have to choose
highest value (instead of minimum value) among all and apply Johnson's rules
with all assumptions same.
The above remarks can be observed in the sequencing problems illustrated
below.
Special Cases In Sequencing : Case (i): Tie in Different Series :
When least values appear at more than one place in different series. (No effect).
ILLUSTRATION 2
Find the sequence for thefollowing data so as to minimise the idle time in hours.
Jobs
A B C D
Machines I 5 6 8 4
II 4 7 9 10
What will be the earliest delivery time that B can be promised.
[JNTU (Mech.) 89/S, (CSE) 92/S]

? lull on :
According to Johnson's rule, with usual assumptions, we sequence four jobs as
follows.
The least time among all the timings is 4 and it appears at two places. One is
in the first series (MI) which is to be kept in extreme left position and second is in
second (Mil) series that comes extreme right.
Sequencing 471
D A

8
M-I

M-II 9
wm
Then we are left with the following
B C

M-II

Least of the above timings is 6 that appears in M-I series, therefore B occupies
next left extreme and obviously the left out position is occupied by C.

Then sequence
D is B C A

Time Chart:
Sequence M-I M-II Idle time
In Process Out In Process Out

D 0 4 4 4 10 14 4

B 4 6 10 14 7 21 -

C 10 8 18 21 9 30 _

A 18 5 23 30 4 34 -

Total elapse = 34, idle time = 4 for M - II

Job B can be delivered at the end of 21st hour of starting the operations.

Case (ii) : Tie in Same Series :

i.e., when lowest values appear at more than one place in the same series.
JLLUSTRATIGfcl 3 — — -
An old shopping complex is intended to be renovated. This is done in two phases
namely repairing the patches etc., and then white washing. The time estimates in
hoursfor each of thesefloorsare given below:
Floors

Repair 20 28 25 20 22
White Wash 24 26 15 27 15
472 Operations Research : Theory and Practice
Find the sequence that minimises the idle time ofrepairing and white washing jobs.
Also find the total elapse. Assume no passing and both come and go at the same time
and find the idle time ofeach jobs. [OU (Stat) 84]

Let us apply Johnsons rule. On observation we find most minimum time at two
places i.e., for third and fifth floors of white washing series (second series), i.e., both
are competing for last position. Thus last two positions should be given to these, but
the question is that in which order these two are to be placed. This tie is resolved by
checking their first series time estimates. These are 25 and 22 respectively. Minimum
of these two is 22 and we know that this should be placed to left if it belongs to first
series. Therefore fifth floor occupies left of the last two positions and third floor
occupies the last position as shown below.

©
Floors 1 2 4

Repair 20 28
iSiilll 20 Ililliliilli
White wash 24 26 27

Next minimum among 1, 2 and 4 floors is 20, that appears twice i.e., for first
and fourth floors in first series making a tie for first position. Therefore they occupy
first two positions and the order can be fixed by seeing their second series.

In second series first and fourth floors have 24 and 27 hours respectively, of
which 24 is least. As this belongs to second series, this should go to right and thus
fourth floor is placed first.

© © © ©
itfipiii
Floors
iiilfii
Repair SB 28

White wash 26

Obviously, the second floor is placed in the left out position i.e., third slot. Thus
the optimal sequence is

Q
Sequencing 473

Elapse Time Calculation :

Sequence Repair White Wash Idle Time

Start Process End Start Process End Repair White
wash
4 0 20 20 20 27 47 20
1 20 20 40 47 24 71
2 40 28 68 71 26 97
5 68 22 90 97 15 112
3 90 25 115 115 15 130 15 3
15 23

As both repairing and white washing are starting and ending at same time, the
repairer will be idle for 15 hours and white washer will be idle for 23 hours.

Total idle time = 15 + 23 = 38 hours.

Total elapsed time =130 hours.

Case (in) & (iv) : Multiple or Alternate Optimal Solution : Tie in both series.

If there is a tie with minimum value in first and second series simultaneously
or if there are identical values in two jobs. [Multiple optimal solutions]
ILLUSTRATION 4
A Desk Top Publishing (DTP) operator has 5Jobs to process for which he he has
ha to
first type a copy of DTP and then printing. The estimated timings (in hours) are
asfollows. Find the optimal sequence in minimise the idle time.

Jobs
A B C D E
DTP 5 6 7 9 6
Printing 5 8 10 11 8

What will be expected delivery time for job C. If there is urgency for C what
sequence is preferred. [JNTU 95/CCC]

Among all the jobs the most minimum is 5 that appears for A in both series
i.e., DTP (first series) and printing (second series).

Since it appears in first series, it should occupy first place (say solution - 1) and
as it is found in second series it can go to last place also (say solution - 2). Thus we find
two solutions as shown below :
474 Operations Research : Theory and Practice

Solution 1 :
A

Solution 2 ;
A

The next minimum value is read as 6 that appears in first series at B and E
places. Therefore, we check their corresponding second series times, but these are
also same. Thus in the first two positions these two can occupy in any order (BE or
EB) resulting in two more alternate solutions.
Solution l(a) ; Solution l(b):

A B E A E B

Solution 2 (a) : Solution 2(b)•'

B E A E B A

Next minimum is 7 with C in first series and so placed left of remaining vacancy
positions and D takes the left out position.
Solution l(a) : Solution l(b) :

A B E C D A E B C D

Solution 2(a) : Solution 2(b) :

B E C D A E B C D A

Time Charts :
For Solution l(a) :

Sequence DTP Printing Idle Time

In Process Out In Process Out DTP Printing
A 0 5 5 5 5 10 - 5

B 5 6 11 11 8 19 - 1

E 11 6 17 19 8 27 - -

C 17 7 24 27 10 37 - -

D 24 9 33 37 11 48 15* -

Total Idle Time 15 6

Sequencing 475

For Solution2 (b) :

Sequence DTP Printing Idle time
In Process Out In Process Out DTP Printing
E 0 6 6 6 8 14 - 6
B 6 6 12 14 8 22 - -

C 12 7 19 22 10 32 - -

D 19 9 28 32 11 43 - -

A 28 5 33 43 5 48 15* -

Total Idle Time 15 6

Total elapse (or completion time of all jobs = 48 hours.

* 15 hours for DTP is not considered as idle time, if DTP operator can be asked
to go after completion of his all jobs, i.e., after 33 hours.

The expected delivery time of C is 32 hours or 37 hours. If there is urgency the

sequence E-B-C-D-A may be followed.

Case (v) : Maximisation Case :

ILLUSTRATION 5 — _ _ _

Mr. Shyam receives Rs. 10/-per hour of processing the jobs. Every job has to
undergo two operations, machining and polishing. The time estimates for four
jobs to beprocessed in this week are given below. What should be the sequence if
he intentionally delays his jobs by maximising the idle time of operators so as to
maximise his returns. What is the amount that he can gain or lose, as compared
to that had to beprocessed with minimum delay.
Jobs
Machining in
(hours)
Publishing in
(hours)

The above problem has two cases, viz case (i) maximisation (ii) minimisation.
Case (i) : Maximisation i.e., when Mr. Shyam intentionally delays the process. To find
the sequence that maximises the idle time, we choose highest among all the
timings of the jobs of both operations and if this falls in first series, place it left
otherwise right.
476 Operations Research : Theory and Practice
Highest among all is 45 for Q that should occupy extreme left since it is in first
series.
Q
Next highest is 40 for R in second series therefore goes last.
Q R

Next highest is 35 for P in second series and so goes right side. And the left out
position is occupied by S. The sequence is
Q s p R

Schedule is as Follows :
Sequence Machining Polishing Idle time
In Process Out In Process Out Machining Polishing
Q 0 45 45 45 15 60 _ 45
s 45 25 70 70 30 100 - 10
p 70 10 80 100 35 135 - -
R 80 20 100 135 40 175 75* _

Total Idel75time 55

Total elapse is 175 hours.

Case (ii) : If he wants to minimise.
We choose least for sequencing by Johnsons rule.
The least is 10 for P in first series, so goes to extreme left and next least is 15
for £) in second series, thus goes right. Next least is for R with 20 in first series to
occupy second position and the rest is for 5.
p R S Q
Now the time chart for this sequence is as follows :
Sequence Machining Polishing Idle time
In Process Out In Process Out Machining Polishing
:
P 0 10 10 10 35 45 — 10
R 10 20 30 45 40 85 _ _
S 30 25 55 85 30 115 _ _

Q 55 45 100 115 15 130 30* _

Total Idle30time 10

Total elapse = 130

Difference in the total elapse = 175 - 130 = 45 hours.
Sequencing 477

Job Total elapse of each job Returns @ Rs 10 per hr Difference

Maximise case Minimise case Maximise case Minimise case Rs

P 135 45 1350 450 900

Q 60 130 600 1300 -700

R 175 85 1750 850 900

S 100 115 1000 1150 -150

Total 470 375 4700 3750 950

Remarks :
•* Assumed that machining operator can go after completmg all his jobs and
hence no idle time/waiting time till polishing is completed.
** The sequence of maximisation is not the reverse order of minimisation case.

Practice Problems
1. Find the sequence that minimizes the total elapsed time (in hours) required to
complete on the following two machines.

Task A B C D E F G

Machine I 2 5 4 9 6 2 7
Machine II 6 8 7 4 3 9 3

[SVU - (M.Sc.) 83, AU-BE (Elect) 83, JNTU (CSIT) 2000/S

JNTU (Mech. & ECE) 2001/CCC]

Answer: Sequence in FACDEG

2. Find the sequence that minimises the total elapsed time (in days) required to
repair and test the following seven jobs.
Activity 1 2 3 4 5 6 7
Repairing time in days 3 5 9 9 6 3 3

Testing time in days 4 5 1 6 3 3 4

[JNTU (Hech.) 88, 0U - 86]

Answer : Sequences 3-1-7-6-2-4-5 or 3-1-7-2-4-6 or

3-7-1.6-2-4-5 or 3-7-1-2-4-6-5
478 Operations Research : Theory and Practice

3. There are 7 jobs each of which has to pass through machines My andM 2 in
order, My -> M2- Processing times (in hours) are given below. Determine the
sequence that will minimise the elapsed time and prepare the time schedule
chart and Gantt chart.

Job 1 2 3 4 5 6 7
Ml 3 15 12 6 10 11 9
M2 8 10 10 6 12 1 3
[JNTU (Mech.) 98]

Answer: Optimal Sequence i s : l - 4 - 5 - 3 - 2 - 7 - 6 or 1-5-3-2-4-7-6

Ml M2
JOB Time in Time out Time in Time out

1 0 3 3 11
4 3 9 11 17
5 9 19 19 31
3 19 34 34 44
2 34 46 46 56
7 46 55 56 59
6 55 66 66 67

Total time elapse = 67 hours.

Idle time on Ml = 1 hour.
Idle time on M2 =17 hours.

4. In a pipe fabrication shop the times required for cutting and bending operation
for 6 different sizes of pipes are given. Determine the order in which those
pipes should be processed in order to minimise the total time for processing
all the pipes.

Pipe 1 2 3 4 5 6
Cutting Time 3 12 5 2 9 11
Welding Time 8 10 9 6 3 1
[JNTU (Mech.) 96/P/S]

Answer: Optimal Sequence i s : 4 — 1 — 3 — 2 — 5 — 6

Sequencing 479

Ml M2
JOB Time in Time out Time in Time out
4 0 9 2 8
1 2 5 8 16
3 5 10 16 25
2 10 22 25 35
5 22 31 35 38
6 31 42 42 43
Total time elapse = 43 hours.
Idle time on Ml = 1 hour.
Idle time on M2 = '6 hours.

5. The repairing activity on eight machines take the time as shown below.
Sequence them to minimise the idle time of repairmen and schedule in a chart.
Machine A B C D E F G H
Repair team - 1 5 4 22 16 15 11 9 4
Repair team - 2 6 10 12 8 20 7 2 21

[JNTU (CSE) 96/S]

Answer: H —» B —>A ; min. Total time = 90 hi s.

6. The time required in days for reparing and overhauling six machines is shown
in the following table. Also, given that overhauling is to be done only after
repairing job. Presently, the following schedule is choosen.
Machine sequence My M2 M3 M4 M
5 M6
Repair time 6 6 4 6 5 8
Overhauling time 4 2 10 6 3 6

(i) Determine optimal sequence

(ii) If downtime of any of the six machines costs Rs. 800 per day, idle time of
repairmen costs Rs. 250 per day and idle time for overhauling costs Rs.
430 per day, which of the two schedules, the present one and the one
found in (i), will be more economical. What are their respective cots.
[JNTU (CSE) -96]

Answer: (i) M 3 -> M6 -> M4 -> Mx -> M5 - M 2 ; (ii) Present costs = Rs. 64, 630,
Optimal cost = Rs. 60,310.
480 Operations Research : Theory and Practice

7. Find the optimal sequence and min. eplased time for the following data.
Job 1 2 3 4 5
Machine P 5 1 9 3 10
Machine Q 2 6 7 8 4

[JNTU (Mech - ECE) 96/C]

Answer: 2 - 4 - 3 - 5 - 1 ; 30 hrs, Idle time on^4 = 2 hrs, B = 3 hrs

8. Ram & Co plans to fill six positions for which it conducts two types of tests. The
experts who came to test the candidates have approximated their probable time
and given in minutes of time as follows. Also it is decided to test aptitute first
followed by interview. Find the sequence that can minimise the waiting time of
the experts.
Position I II III IV V VI
Aptitude test 140 180 150 200 7 100
Interview 70 120 110 80 100 90

Answer: II - III - V - VI - IV - I ; Min time = 440 min.

9. The Hi-tech Publishers have to publish 8 text books for which it spends the
time in days for DTP preperation and printing are estimated as follows, find
the best sequence to take up these books to minimise total publishing time and
also find the idle time of DTP operator and printing operator.
Books B2 *3 B4 *5 5
6 B7 B
S
DTP 14 26 17 11 9 26 18 15
Printing 21 15 16 21 22 12 13 25
A.'ns'wer .* JJF ~^ISA ——^ -MI —^ -^5ft ~~^ &% ~~^ - ^ 9 ™~^ *JH —^ *^c

Optimal publishing time =154 days,

Idle time for DTP operator = 18 hrs for printing = 9 hrs.
10. Find the sequence that minimises the total elapsed time (in hours) required to
complete following tasks on two machines :
Task 1 2 3 4 5 6 7 8 9
Machine A 2 4
Machine B 6 11
[JNTU (Mech. & ECE) 98/CCC, CSE/ECE 2001]
Answer: l-»9-*3->2->8->5-*4-»5-»7or
1^9^3_>8-^2^6^4->5-^7
Minimum time is 61 hours, Idle time on MA = 1 1 hrs,
on Mg = 2 hrs.
Sequencing 481
11. A company has six works on hand, coded^4 to F. All the works have to go through
two machines Mx and MI}. The time required for works on each machine, in
hours, is given below :

Work G H I J K L

Machine I 1 4 6 3 5 2

Machine II 3 6 8 8 1 5

Draw a sequence table scheduling the six works on the two machines.
[JNTU (Mech.) 96/P/S]
Answer: G -» L ->J -^ / / - > / - > K
Minimum time is 32 hours.
12. A company has 3 works on hand. Each of these must be passed through two
departments, the sequential order for which is :
Department A : Press shop Department B : Finishing

The table below lists the number of days required by each job in each
department

Work 1 Work 11 Work 111

Department A : 8 6 5
Department B : 8 3 4

Find the sequence in which the 3 works should be processed so as to take

minimum time to finish all the 3 works.
Answer; Work I —> work III —> work I I ; Minimum time is 23 days.

13. Six jobs go fiiM o T machine I and then over machine II. The order of the
completion of job h.<> no significance. The following table gives the machine
times in hours for six jobs as the two machines.

Job A B C D E F
Time on machine I 5 9 4 7 8 6
Time on machine II 7 4 8 3 9 5

Find the sequence ofjobs that minimise the total elapsed time to complete the
jobs. Find the minimum time by using Gantt's chart or by any other method.
[JNTU (CSE) 92/S]
Answer; C-^A-^E^F-^B-^D; Minimum time is 42 hours.
482 Operations Research : Theory and Practice

14. Five jobs are to be processed on two machines. Machining times are given in
the following table :

Machines/Jobs Jx h h y5
10 5 15 22 3

M2 6 8 12 4 15

Schedule the jobs on machines. [JNT B.Tech. (Hech.) 2001]

[JNTU

Answer : Jb - > / 2 - > / 3 - » / i ->/4 ; Elapse = 59 hrs.

15. Solve the following 2M/C x n job sequencing.

Job 1 2 3 4 5

M- 1 3 7 4 5 7
M-2 6 2 7 3 4
[JNTU (CSE) 96]

Answer : 1 - 3 - 5 - 4 - 2 ; Min. time = 28 hrs;

idle time 2 hrs on Mj and 6 hrs on M 9

16. A book binder has to print and bind six books, the binding times required in
minutes are 80, 100, 90, 60, 30 and 10 while that for printing are 30, 120, 50,
20, 90 and 110 min. respectively. Find optimum sequence, minimum elapsed
time and idle time for each operator.
Answer : Sequence 4 —> 1^3—» 2 —» 5 —» 6, Min. Elapse = 430 min.
Idle time for printer = 10 min, binder = 40 min.
17. Solve the sequencing problem for six jobs whose process times in hrs on
machined, Machine B are (4, 6), (8, 3), (3, 7), (6, 2), (7, 8) and (5, 4).
Answer : 3 - 1 - 5 - 6 - 2 - 4 ; Total Elapse = 35 hrs

18. Determine the sequence and minimum -elapse for the following process times
in hrs.

Job 1 2 3 4 5
M/CP 10 2 18 6 20
M/Cg 4 12 14 16 8
[JNTU (Mech.) 93/P/S]

Answer: 2 - 4 - 3 - 5 - 1 ; Total elapse = 60 hrs.

Sequencing 483

19. In factory, there are six jobs to perform each of which should go through two
machines^ and B, in the orders, B. The processing timings (in hours) for the
jobs are given below. You are required to determine the sequence for
performing the jobs that would minimize the total elapsed time T. What is the
value of 7?
Job Machine A Machine B
h 1 5

h 3 6

h 8 3

h 5 2

h 6 2

h 3 10
[JNTU Mech/Prod/Chem 2001/S]

9*3 Sequencing n Jobs X 3 Machines: Jackson's Rule

For sequencing and scheduling n jobs on three machines, Jackson proposed a
simple rule by which the n X 3 is converted to n X 2 model for preparing the sequence.
And then schedule may be prepared as usually for the three machines individually
either by time chart or Gantt chart.

For converting the n X 3 to n X 2 the given problem has to satisfy one or both
of the following conditions.
Condition 1 :
The minimum process time among all the jobs to be performed on first machine
is greater than or equal to the maximum process time among the jobs to be performed
on second machine, i.e., Min (MY) > Max (Af2)
Condition 2 :
The minimum process time among all the jobs to be performed on third
machine is greater than or equal to the'maximum process time among the jobs to be
performed on second machine. Min (M3) > Max (M2)
(In other words the maximum of second machine is less than or equal to the
minimum of first or/and third machine timings). Max (M2) < Min (Mx or/and M3)
If the problem satisfies one or both of the above conditions, the 3 X n
sequencing problem is converted to 2Xn sequencing problem by adding the job
timings of first machine with the corresponding job timings of second machine to
make as first center and similarly adding job timings of second machine with their
respective job timings of third machine to make it as second work center. Now, the
sequencing is made for the two hypothetical work centers with these new timings
according to Johnson's rule.
484 Operations Research : Theory and Practice

i.e., WCX = (My + M2j) and WC2 = (M2j + M*j)

After finding the sequence, the Gantt chart and time chart calculations are done
by considering three machines with the timings given in the original problem.

A numerical example is shown here below to illustrate

ILLUSTRATION. 6 —•

A company uses its maintenance crew in three teams for their preventive
maintenance of their heavy vehicles. First team looks after the replacement of
worn out parts, the second oiling and resetting and the third checking and tests
running. The estimated time for maintenance of each of these vehicles is given in
hours in the following table and passing is not allowed. Find the sequence and
schedule them so as to minimise the total elapsed time and idle time.
Team/Vehicle No. 1 2 3 4 5 6 7
Replacement team 3 8 7 4 9 8 7
Resetting team 4 3 2 5 1 4 3
Inspection team 6 7 5 11 5 6 12
[JNTU-(Mech.)2001/C/S]

Seven vehicles are given to be processed through three repair teams say
i?1? R2, #3- In order R^, R2, ^s- Now, the minimum time in i?j series is 3, and that in.
R$ is 5 while the maximum in R2 ls 5- Since minimum i?3 > maximum R2 (condition
2) is satisfied; Hence the problem can be converted to that of 7 jobs X 2 work centers.
Thus WCX =RX + Rc, and WC2 = R2 + R$

The problem is re-written as

WC/V.No 1 2 3 4 5 6 7

WC1 3+4=7 8 + 3 = 11 7+ 2=9 4+ 5=9 9 + 1 = 10 8 + 4 = 12 7 + 3 = 10"

WC2 4 + 6 - 1 0 3 + 7=10 2 + 5 = 7 5 + 11 = 16 1+5=6 4 + 6=10 3 + 12= 15

Using Johnson's optimal sequence algorithm, the correct sequence is*

(* Students are advised to write sequence step by step as shown in the previous illustrations.)

For calculating the total time elapsed and the idle times, we draw the time chart
as follows
Sequencing 485

Team Replacement Team Resetting Team Inspection Team Idle Time / Waiting
Vehicle Timings (Rx) Timings (i?2) Timings (£3) Time
No.
In Repair Out In Repair Out In Repair Out R,
V\ 0 3 3 3 4 7 7 6 13 - 3 7

3 4 7 7 5 12 13 11 24 - - -
V
7 7 7 14 14 3 17 24 12 36 - 2 -
v2 14 8 22 22 3 25 36 7 43 - 5 -

22 8 30 30 4 34 43 6 49 - 5 -
30 7 37 37 2 39 49 5 54 - 3 -
37 9 46 46 1 47 54 5 59 13 7+12 -
Total 46 46 22 47 52 59 13 + 37 + 7 = 57

The table indicates that the total elapsed time is 59 hours. Idle time of
replacement team is 13 hours at the end of project. Resetting team gets 37 hours of
idle time of which are 3 hours at the starting and 12 hours at the ending and 22 hours
in middle of the project. The inspection team will have to wait 7 hour at the beginning
of the project.
Gantt Chart:

Idle time of inspection

/ 7 13 24 36 43 49 54 59
Inspection
© V.

(6) (12) (7) (6) (5) (5)

30 34 137 39 | 46 47 Idle times
of resetting
Resetting

(4) (5)
Idle time
of repair
Repair
i
(3) (4) (7) (8) (8) (7) (9)

-I h-
10 15 20 25 30 35 40 45 50 55 60

FIGURE 9,2: GANTT CHART

486 Operations Research : Theory and Practice

ILLUSTRATION 7
Find the sequence that minimizes total machining time to complete the
following data:
Tasks A B C D E F
Time on Machine I 4 9 8 5 10 9
Time on Machine II 5 4 3 6 2 5
Time on Machine III 7 8 6 12 6 7

[JNTU (CSE) 97]

Solution !
Conditions : Min. of Mj > Max. of M 2 or Min of M 3 > Max. of M 2
Here 6 = 6 , Second condition of Jackson's rule of sequencing algorithm is
satisified. Hence we can conver£the above 3 machines Xn jobs case to 2 machines
Xn jobs case.
Rearranging the Table :
Tasks A B C D E F
Mx+M2 9 13 11 11 12 14

M2 + M3 12 12 9 18 8 12
st
Note : Identify the least processing time of all the values. If this falls in I machine, allocate the task at
left and if it falls in 217 machine allocate to right Repeat the process till all the tasks are scheduled.
(Students are advised to show all steps)

Optimal Sequence is : A -» D -> B -

Machines M1 M2 Idle Time

Job
Time Process Time Time Process Time Time Process Time Ml M2
in Time out in Time out in Time out
A 0 4 4 4 5 9 9 7 16 — 4 9
D 4 5 9 9 6 15 16 12 28 — — _
B 9 9 18 18 4 22 28 8 36 — 3 _
F 18 9 27 27 5 32 36 7 43 — 5 _
C 27 8 35 35 3 38 43 6 49 — 3 —
E 35 10 45 45 2 47 49 6 55 10 7 + 8 -
Total 55 10 30 9
Total elapsed time = 55 hrs ; Idle time on m/c 1 = 10 hrs.
Idle time on m/c 2 = 30 hrs. and
Idle time on m/c 3 = 9 hrs.
Sequencing 487

ILLUSTRATION 8

A company has 6jobs which go through 3 machines X> YandZin the order XYZ.
The processing time in minutesfor eachjob on each machine is asfollows:
Job
3
Machines X 18 12 29 36 43 37
Y 7 12 11 2 6 12
19 12 23 47 28 36

Find the optimal sequence, total elapsed time and idle timesfor each machine.
[JNTU (Mech.) 99]

Solution :
Since xm[n > 3>max, we have

1 2 3 4 5 6
A X+Y 25 24 40 38 49 49
Y+Z 26 24 34 49 34 48

The correct sequence is

h ->/i ->/4 ->7e ->/3 or

/i

Machine X Machine Y Machine Z

Job Initial Process Final Idle Initial Process Final Idle Initial Process Final Idle
Time Time Time Time Time Time Time Time Time Time Time Time
2 0 12 12 — 12 12 24 12 24 12 36 24
1 12 18 30 — 30 7 37 6 37 19 56 1
4 30 36 66 _ 66 2 68 29 68 47 115 12
6 66 37 103 _ 103 12 115 35 115 36 151 0
3 103 29 132 132 11 143 17 151 23 174 0
5 132 43 175 34 175 6 181 32 + 181 28 209 7
28

(i) Total clapped time = 209 minutes

(ii) Idle time on machine X = (209 - 175) = 34 minutes

(hi) Idle time on machine Y = 12 + 6 + 29 + 35 + 17 + 32 + (209 - 181)

= 159 minutes
(iv) Idle time on machine Z = 2 4 + 1 + 12 + 7 = 44 minutes.
488 Operations Research : Theory and Practice

ILLUSTRATION 9

There are 5 projects to be made, each of which must undergo three stages viz,
information feeding, analysing and printing. A programmer has decided to
operate three computers namely Zebronics, Wipro and Acer for these
operations. Theprinter is connected to Zebronics only.
Processing times are given below:
Project No. 1 2 3 4 5
Acer 8 5 4 6 5
Zebronics 10 13 11 10 12
Wipro 6 2 9 7 4

(i) Determine the total elapsed time and idle time ofeach computer.
(ii) If it can be processed by the sub-contract services by three out side parties,
NUT, CMC and STG to process on Zebronics, Wipro and Acer
respectively, schedule the parties optimally, under the condition that the
parties may be called on any day but the contract once started should be
continued till the last job of the respective party is completed and the
payment should be madefor the waiting idle times also.
(Hi) What will be the amounts paid to each party if it costs Rs. 10/-for each
working hour andRs. 5/-for each waiting hour.
(iv) If any project that waits will cost Re.l/- per hours, what will be the cost
incurred on waiting of projects. Assume all projects are taken up now and
waiting time is counted from the time it is taken up.
Solution
As per the given data, the.printing is the last operation and is connected to
Zebronics only. Therefore process times on Zebronics should be taken last i.e., as
machine - 3, and the other two can be taken in any order if they satisfy Jackson's
conditions.
Now let us examine which order of machines is applicable here to convert this
n jobs X3 machines case to n jobs X2 machines case.
Project 1 2 3 4 5
Acer 8 5 4 6 5
Wipro 6 2 9 1 4
Zebronics 10 13 11 8 12

Case (i) : Order of machines as Acer -» Wipro -» Zebronics

Condition Checking :
Min. of Acer times (A{) : 5 — (Machine - 1)
Min. of Zebronics times (Zj) : 8 — (Machine - 3)
Max. of Wipro times (WJ) : 9 — (Machine - 2)
Sequencing 489

Here Min. of Zebronics times (Zr) £ Max of Wipro {W0 and Min. of Acer
(Ai) £ Max. of Wipro (W$. i.e., 5, 8 < 9.

Both the Jackson's conditions are not satisfied in this case and hence we cannot
proceed with this order.
However, let us check the condition for case (ii)

Case (ii) : Order of machine as Wipro —> Acer —> Zebronics.

Condition Checking :
Min. of Wipro (W$ : 2 — (Machine - 1)
Min, of Zebronics (Z-) : 8 — (Machine - 3)
Max. of Acer (Aj) : 8 — (Machine - 2)
Here Jackson's condition (ii) is satisfied as
Min. of Zebronics (Z^) > Max. of Acer (A$
i.e., 8=8
Therefore we go by the order of the machines as
Wipro —> Acer —> Zebronics

Project 1 2 3 4 5
Wipro 6 2 9 7 4
Acer 8 5 4 6 5
Zebronics 10 13 11 8 12

Since, the second condition of Jackson's rule is satisfied we can convert the
above n jobs X 3 M/C model to n jobs X 2 M/C model to find the optimal sequence.
This is done by taking two hypothetical machines say X and Ywhere the process times
are given by

Xi=Wi+Ai and Y^A^

Thus the problem reduces to
Project 3 4 5
Machine X 13 13 9
Machine Y 18 |^#:> 15 14 17

Now, the least among all the ten timings is 7 that appears in first series for
project - 2. Therefore this is placed extreme left and project 2 is deleted for the next
iteration.
490 Operations Research : Theory and Practice

Project
Machine X 14 13 13
Machine Y 18 15 14

Now, least is 9 in first series for project - 5, and so occupies left side i.e., second
place, and project - 5 is deleted for next iteration.
(?) (5)

Project
Machine X 14
Machine Y 18

Next least timing (13) is in tie between project - 3 and 4, both in first series for
next two positions. Therefore, we see second series. The least (14) appears for project
- 4, therefore it goes right side of the third and fourth position. The sequence will be

Obviously project - 1 occupies the last position.

Thus the optimal sequence is 2 —» 5 -» 3 —» 4 —> 1.

The scheduling chart for the above sequence is calculated as follows.

WIPRO ACER Waiting Time For

ZEBRONICS IDLE TIMES
Projects
Sequence
In Proc Out In Proc Out In Proc Out W A Z At At At
Total
W A Z
Proj. - 2 0 2 2 2 5 7 7 13 20 *_ NIL NIL NIL
Proj. - 5 2 4 6 7 5 12 20 12 32 - - - 2 1 8 11
Proj. - 3 6 9 15 15 4 19 ' 32 11 43 3 _ 6 _ 13 19
Proj. - 4 15 7 22 22 6 28 43 8 51 3 15 15 30
Proj. - 1 2.2 6 28 28 8 36 51 10 61 *_ *_ - 22 . 15 37

Elapse time 28 28 28 36-2 54 61-7

- 6 - 45 1 51 97
= 34 = 54

As we can call the parties at any time, it is better to call CMC first to work on
Wipro from 0 to 28 hours and STG to work from second hour to 36th hours on
Acer and NIIT to start at 7th hour to finish at 61st hour on Zebronics.
Sequencing 491
The cost calculations are
Party Computer Cost of working Cost of waiting Total cost
CMC WIPRO (0 to 28) 28 x 10 = 280 NIL 280
STG ACER (2 to 36) 2 8 x 1 0 = 280 6 x 5 = 30 310
NUT ZEBRONICS(7to61) 54 x 10 = 540 NIL 540
* 28 hrs actual working and 6 hrs; Total 34 for ACER Total cost (Rs) 1130

Waiting cost of projects : 97 x 1 = Rs. 97/-

9.4 Sequencing n Jobs X m Machines

Sequencing the jobs may not be always in the form of n X 2 or n X 3. In fact in
real practice, we often come across with the n jobs to sequence on m machines or men
or work centers. However it is not very difficult to sequence in such case. The method
we apply here is again based on the Jackson's conditions as explained below.

Let there be n jobs, of which each is to be processed through m work

centers/machines say 14^, W%, W$, . . . Wm.

Now, the iterative procedure is as follows:

Step 1: Find minimum process time of Wx and Wm series, and maximum process time
of all the middle series.
Step 2: Check whether

Condition 1 : Min of PFj series > max of middle series.

Condition 2 : Min of Wm series > max of middle series.

Step 3 : If both conditions of step 2 are not satisfied, the method fails. If at least one of
the conditions is true, then go to next step.

Step 4 : Convert m work center problem to two-work center problem by assuming two
fictitious machines say A and B, where the process times on A and B are
Ai = the algebraic sum of process times of W^ and all corresponding timings of
the mediocre series and B • = the algebraic sum of process times of Wm and all
corresponding timings of the mediocre series.
i.e.f Ai=Wlj+W2j+W5j+....W{m_l)j
and Bi=W2j + W3j+W4j+ -...Wmj

(However in general practice we can find the sequence by taking the timings of
W^ and Wm if the sums of the mediocre series are fixed positive constants).
492 Operations Research : Theory and Practice

ILLUSTRATION 10
Five machine parts are to be reassembled after yearly maintenance in the order
P. Q, RySandi on four machines A. B, CandD. Find the optimal sequence for
the machines whenpassing is not allowed of which the repair time (in hours) is
given below. Also find the total elapsed time, and individual idle times for the
repairmen on each machine and machine part (Consider that the machine will
be handed over to the production department immediately after reassembling all
theparts).
Machine/Part P Q R S T
A 7 5 2 3 9
B 6 6 4 5 10
C 5 4 5 6 8
D 8 3 3 2 6
Solution ;
Here the machines are to be sequenced
Checking Sequencing Algorithm Rules :
Minimum repair time of P series [Min. (tP)] - (tPC) = 5
Minimum repair time of T series [Min. (tj;)] = (tjp) - 6
And maximum repair time of Q R, S series [Max ( ^ tR: tsp] = {6, 5, 6}
Since the condition Min. (tjj) > Max (tQj, tRj, tsj) is satisfied the given problem
can be converted into a four jobs and two machines (say X and Y) problem as

Machine/Job A B C D
Machine X 17 21 20 16
Machine Y 19 25 23 14

Now, using the optimal sequence algorithm, the following optimal sequence
can be obtained.
A C B D
The total elapsed time corresponding to the above sequence can be computed
as shown in the following table, using the individual processing times as given in the
original problem.
Machine/Part P Q R S T
A 0-7 7-12 12-14 14-17 17-26
C 7-12 12-16 16-21 21-27 27-35
B 12-18 18-24 24-28 28-33 35-45
D 18-26 26-29 29-32 33-35 45-51
Sequencing 493

From the above table, it can be understood that the minimum total elapsed
time is 51 hours.

The idle times for the man who reassembles the machine parts are as follows :

For P = 51 - 2 6 = 25 hours.
For Q = 7 + (18 - 16) + (26 - 24) + (51 - 29) = 33hours.
For R = 12 + (16 - 14) + (24 - 21) + (29- 28) + (51 - 32) = 37 hours.
For S = 14 + (21 - 17) + (28 - 27) + (51 - 35) = 35 hours.
For T = 17 + (27 - 26) = 18 hours.

The waiting times for the machines are as follows:

For A = No waiting time and will be handed over after 26 hours.

For B = 12 + (35 — 33) = 14 waiting hours and will be handed over after
45 hours.
For C = 7 waiting hours and will be handed over after 35 hours.
For D = 18 + (33 - 32) + (45 - 35) = 29 waiting hours and will be handed
over after 51 hours.

Practice Problems
1. A mechanic has to assemble 5 machines and each assembly takes three stages
for which he employed three persons. The sequence of the stages is fixed as
stage I, stage II and stage III. Determine the optimal sequence for the machines
to be assembled to minimise the time elapsed from the start of first machine to
the end of last machine. The time required, in days for each machine at each
stage is given in the following matrix. Also find the time that the mechanic could
promise to deliver each machine. If the mechanic can delink a person from the
present process if his stage work is completed or not started, what will be the
net idle time that the process demands?

Machine Code Stage I Stage II Stage III

M0235988 6 8 14
K0242115 16 10 18
V6566573 14 2 10
G3042721 10 4 12
S6567763 8 6 20
494 Operations Research : Theory and Practice

Answer : Optimal Sequence is : M - G- S - K-V

Machine Stage -1 Stage - II Stage - III Delivery

day
Time in Time out Time in Time out Time in Time out From now
M 0 6 6 14 14 28 28
G 6 16 16 20 28 40 40
r

S 16 24 24 30 40 60 60

K 24 40 40 50 60 78 78
V 40 54 54 56 78 88 88

Total elapsed time is 88 hours and idle time for stage -1 - 34 hours (the mechanic
can delink all the last 34 hours), for stage - II - 58 hours (in which 6+ 32 = 38
hours can be delinked) and for stage III - 14 hours (in which all the 14 can be
delinked). Net idle time due to process is 20 (stage II operator) if delinked
2. There are 5 spare parts are to be made each of which must go through machines
Ml, M2, andM3 in order Ml, M3, M2. Processing times are given below.
(i) Determine the optimal sequence, total elapsed time and idle time of each
machine.
(ii) If it can be processed by the sub-contract services by three outside parties,
PI, P2 and P3 to process on Ml, M2 and M3 respectively, schedule the
parties optimally, under the conditions that the parties may be called on
any day but the contract once started should be continued till the last job
of the respective party is completed and the payment should be made for
the process delays also.
(iii) What will be the amounts paid to each party if it costs Rs. 10/- for each
working hour and Rs. 5/- for each waiting hour?
Spare *art 1 2 3 4 5
Ml 8 5 4 6 5
M2 10 13 11 10 12
M3 6 2 9 7 4
[JNTU (Mech.) 96/CCC, (ECE) -96]
Answer: (i) O p t i m a l s e q u e n c e i s : 2 — 5 — 3 — 4— 1
Job Ml M2 M3
Time in Time out Time in Time out Time in Time out

2 0 5 5 7 7 20
Sequencing 495

5 5 10 10 14 20 32
3 10 14 14 23 32 43

4 14 20 23 30 43 53
1 20 28 30 36 53 63

Total elapsed time = 63 hours.

Idle time on Afi = 35 hours, on M2 = (27 + 5) = 32 hrs and onM3 = 7 hrs.

(ii) PI starts at Oth hour and ends at 28th hour (28 hours), P2 starts at 5th
hour and ends at 36th hour (31 hours) and P3 starts at 7th hour and ends
at 63th hour (56 hours).

(iii) Payments for PI - Rs. 280/- (all are working hours), for P2 - Rs, 295/- (28
working hours and 3 waiting hours), and for PS - Rs. 560/- (all working
hours).
3. Find the sequence that minimises the total elapsed time in hours to complete
the following jobs on 3 machines. Prepare Gantt chart.
A B C D E
Ml 3 8 7 5 2
M2 3 4 2 1 5
M3 5 8 10 7 6
[JNTU (Mech.) 87]

Answer: Optimal Sequence is :A - D - E - C - B or D -A - E - C - B

JOB Ml M2 M3
Time in Time out Time in Time out Time in Time out

A 0 3 3 6 6 11
D 3 8 8 9 11 18
E 8 10 10 15 18 24
C 10 17 17 19 24 34
B 17 25 25 29 34 42

Total elapsed time = 42; Idle time on Ml = 17 hours.

Idle time on M2 = 27 hours.
Idle time on M3 = 6 hours.
496 Operations Research : Theory and Practice

4. Find the sequence that minimises total machining time. The Machine sequence
isPRQ
Tasks A B C D £
Time on M/C - pP 4 9 8 5 10 9
Time on M/C - 7 8 6 12 6 7
R
Time on M/C - R 5 4 3 6 2 5
[JNTU (Mech. & ECE) 97/CCC/S]
Answer: Optimal Sequence is : A - D -B-F-C- E
JOB P R Q
Time in Time out Time in Time out Time in Time out
A 0 4 4 9 9 16
D 4 9 9 15 16 28
B 9 18 18 22 28 36
F 18 27 27 32 36 43
C 27 35 35 38 43 49
E 35 45 45 47 49 55
Total elapsed time = 55 hours. ; Idle time on M/C P = 10 hours.
Idle time on M/C R = 30 hours. Idle time on M/C Q = 9 hours.
5. Find the sequence that minimises total elapsed time if the order of machines

Tasks A B C D E F G
Time on M - 1 3 8 7 4 9 8 7
Time on M - 2 6 7 5 11 5 6 12
Time on M - 3 4 3 2 5 1 4 3
[JNTU (Mech. & ECE) 2001/C/S]
^4«5^er: Optimal Sequence is : A - D --G-B-F--C-E
JOB Ml M3 M2
Time in Time out Time in Time out Time in Time out
A 0 3 3 7 7 13
D 3 7 7 12 13 24
G 7 14 14 17 24 36
B 14 22 22 25 36 43
F 22 30 30 34 43 49
C 30 37 37 39 49 54
E 37 46 46 47 54 59
Total elapsed time = 59 hours.; Idle time on Ml = 13 hours.
Idle time on M2 = 7 hours. Idle time on M3 = 37 hours.
Sequencing 497
6. We have six jobs, each of which must go through machines P, QandR in the
order PQR. Processing time (in hours) are given in the following table.
Job 1 2 3 4 5 6
Machine P (P{) 8 3 7 2 5 1

Machine g (Q) 3 4 5 2 1 6
Machine R (R$ 8 7 6 9 10 9

Determine a squence for the five jobs that will minimise the elapsed time T.
[JNTU - CSE 97]

Answer : Optimal sequences :(i) 4 - 5 - 6 - 2 - 1 - 3 (ii) 4 - 5 - 6 - 2 - 3 - 1

Idle time is 27 hr. for machine P, 32 hr. for machine Q and
4 hr. for machine R. Total ellapsed time = 53.
7. Find the sequence that minimises the total elapsed time required to complete
the following tasks. Each job is processed in the order I - II - III.
Job A B C D E F G
Machine I
Machine II
Machine III
Answer: Optimal sequence \B-F-C-A-E-D-G [JNTU - Mech 97/P/S]
Elapsed time : 59 hours
Idle time is 7 hours for machine I, 13 hours for machines B and 37 hours
for machine III.
8. There are five jobs, each of which must go through machines A, B and C in the
order ABC. Processing times (in hours) are given in the following table.
Job 1 2 3 4 5
Machine A
Machine B
Machine C
Answer: 3 - 2 - 4 - 1 - 5 ; minimum time is 51 hours [JNTU - CSIT 2000]
9. Determine the optimal sequence and idle time of each mahine that will
minimise the total elapsed time.
Job No. 1 2 3 4 5
Machine A 5 7 6 9 5
Machine B 2 1 4 5 3
Machine C 3 7 5 6 7
[JNTU - CSE - 2000]

Answer. 2 - 5 - 4 - 3 - 1 or 5 - 4 - 3 - 2 - 1 or 5 - 2 - 4 - 3 - 1
Elapsed time = 4- hours; Idle time is 8 hrs on machined, 25 hrs on
Machine B and 12 hrs on Machine C,
498 Operations Research : Theory and Practice

10. Give the following data

(a)
Job 1 1 2 3 4 5 6
Machine A 12 10 9 14 7 9
Machine B 7 6 6 5 4 4

Machine C 6 5 6 4 2 4

(b) Order of the processing of each job : ACB

(c) Sequence suggested : Jobs 5, 3, 6, 2, 1, 4

(i) Find the total time elapsed for the sequence suggested?

(ii) Is the given sequence optimal?

(iii) If your answer to (ii) is 'No', find the optimal sequence and the total
elapsed time associated with it. [JNTU CSE 2001/S]

Answer : (i) Total elapse = 70 hrs, Idle times on machined = 9; Machine B = 38;
Machine C = 43
(ii) No.
(iii) Optimal sequene is 1 - 3 - 2 - 4 - 6 - 5
Idle time on Machined = 6, on Machine B = 35 and on
Machine C = 40 hours
11. Find an optimal sequence for the following sequencing problem of four works
and five machines when passing is not allowed. Processing time (in hours) is
given below;

Work A B C D
Machine Ml 6 5 4 7

Machine M2 4 5 3 2

Machine M3 1 3 4 2

Machine M4 2 4 5 1

Machine M5 8 9 7 5

Also fmd the total elapsed time.

Answer : Optimal sequence : A - C - B - D;
Minimum total elapsed time is 43 hrs.
Sequencing 499

• • • 9.5 Sequencing of2 Jobs X n Machines

Suppose there are two jobs say I and II to be processed on V machines say
M} , M 2 . . . Mn in two different orders. The technological sequence of the operations
on the two jobs is already fixed, but the exact or expected time of processing time is
not known. Now, the objective of the problem is to determine the optimal sequence
of processing that results in minimum idle time. In such case's, a graphical solution is
most suitable by taking the two jobs on the two co-ordinate axes. This method is
summarised in the following steps.

Algorithm :

Step 1: Draw the perpendicular axes, such that x-axis represents Job - 1 and y-axis
represents Job - 2. Thus we can say that when process moves horizontally along
x-axis, Job -1 is under process while Job -2 remains idle. And if we move along
y-axi$ (i.e., vertically), Job - 2 will be under process while Job - 1 is idle. If we
move along 45° to either axis, both jobs are simultaneously processed.

Step 2 : Mark the technological sequences and the process times on the axes as given.
Step 3 : Construct various blocks starting from origin, to represent area of the machine
that will be common to both jobs in first quadrant. These are the areas at which
both jobs can not be performed simultaneously. Mark the end point i.e., farthest
corner of last block.

Step 4 : Draw the straight line starting from origin in 45° angle to either of axis. When
the line is obstructed by a block, go along the side of block either horizontally
or vertically as can be moved. Continue the line till end point is reached.

Step 5: The diagonal line (at 45°) represents that both jobs are simultaneously processed,
vertical line represents that Job-1 is idle while Job-2 is under process and
horizontal line represents that Job-2 is idle while Job-1 is under process. Thus,
choose the path in which diagonal movement is maximum. This is said to be
optimal path.
Step- 6: Calculate the total horizontal distance (idle time of Job-2) and total vertical
distance (idle time of Job-1) on optimal path. These distance are added to the
process time represented by the co-ordinates of end of point, as follows.
Total Elapsed Time of Job-1 :
= value of x-co-ordinate of end point + total vertical distance on optimal path
= process time of Job 1 + idle time of Job-1.
Total Elapsed Time for Job-2 :
= value of y co-ordinate of end point + total horizontal distance on optimal
path.
= process time ofjob 2 + idle time of job 2.
500 Operations Research : Theory and Practice

The above procedure is illustrated through an example given below :

ILLUSTRATION 11 — — _ _ _ _ _
There are twojob&to be processed through
d J five machines A, B, C,DandE. The
prescribed technologicalorder is
Job-1: A^B^C^D^E
Job-2: B^C^A^D^E
Theprocess times in hours are given below.
Machine A B C D E
Job - 1 3 4 2 6 2
Job - 2 3 5 4 2 6

Find out the optimal sequencing of thejobs on machines and the minimum time
required to process these jobs.
So\i\\\ov\ i
The above problem is in 2 jobs X n machine model for which we use graphical
method. Let us take Job-1 on x-axis and Job -2 on y-axis and represent the process
time in their technological sequence.

FIGURE 9.3 ;

Now, we prepare the blocks that are common to both Job 1 and 2. Then we
find a diagonal line at 45° starting from origin to reach the end point. The block B
obstructs after moving to (3, 3) units and therefore we move vertically for 2 units and
then go by 45° line. At (9, 11) we can get two paths.One of them is to go 3 units
vertically from (9, 11) to (9, 14) and again at 45° from (9, 14) to (15, 20), then go 2
units horizontally [(15, 20) to (17, 20)] to reach the end point. This gives 2 units
horizontal and (2 + 3) units vertical distances.
Sequencing 501
The second path is along (9, 11) -> (15, 12) -» (17, 14) -> (17, 20). This path
gives a total of 8 (2 + 6) units of vertical and 5 units of horizontal distance. These idle
times are more than those in first path, therefore first path is optimal.
The calculation of elapsed time
For Job-1 is
Process time + idle time
= x-coordinate of end point + total vertical distance on optimal path.
•= 17+ ( 2 + 3) = 22 hours.
For Job 2 :
Process time + idle time
= ^-coordinate of end point + total horizontal distance on optimal path.
= 20 + 2 = 22 hours.
Remark:
If we go by second path, the elapse would have been 17 + 8 = 25 for Job - 1
and 20 + 5 = 25 for Job - 2 which are not optimal.

Practice Problems
1. Use the graphical method to find the minimum elapsed time sequence of 2
works and 5 machines when we are given the following information.
Workl Sequence A B C D E
Time (hours) 2 3 4 6 2
Work 2 Sequence C A D E B
Time (hours) 4 5 3 2 6
[JNTU (Mech.) 99/P/S]

Answer: Idle time is 3 hours for work 1 and zero hours for work 2.
Elapsed time for work 1 is 17 + 3 = 20 hours, for work 2 it is (20 + 0) = 20
2. Two jobs are to be processed on four machines A, B, C andZ> technological
order for these jobs on machines is as follows;
Job A 1 2 3. 4
JobB 4 2 1 3

Processing times are given in the following table

Machines
Jobs 1 2 3 4
Job A 4 6 7 3
JobB 4 7 5 8

Find the optimal sequence ofjobs on each of the machine.

[JNTU (Mech.) 97/P/S]
502 Operations Research : Theory and Practice

Answer: Idle time is 4 hours for job A and zero for Job B.
Elapsed time for Job A is 20 + 4 = 24, for job B it is (24 + 0) = 24.

3. A machine shop has six machines A,B, C, D, EandF. Two jobs must be
processed through each of machines. The times one machines and the
necessary sequence of the jobs through the shop are given below.

Order 1 2 3 4 5 6

Job-I A -20 C - 10 D -10 B-30 £-25 F- 16

Job - II A -10 C- 30 B - 15 D -10 F- 15 £-20
[AU (B.E.) 86]
Answer: Minimum total time = 150 hrs.

4. A machine shop has four machines/I, B, C and D. Two jobs must be processed
through each of these machines. The time and necessary sequency of jobs
through the shop are given below.
Jobl Sequence A B C D

Time (hrs) 2 4 5 1
Job 2 Sequence D B A C

Time (hrs) 6 4 2 3

Use graphic method to obtain the total minimum elapsed time.

[JNTU (Hech.) 95/CCC]

Answer : 15 hrs

5. Solve the following 2 jobs x n machine sequencing problem. Also determine

which jobs is to be taken first on each machine.
Jobl Sequence A B C D E
Time 3 4 2 6 2
(hrs)
Job 2 Sequence B C A D E
Time 5 4 3 2 6
(hrs)
[JNTU (CSE) 97/S, (Mech.) - 98]

Answer : Total time = 22 hrs, idle time for Job 1 = (2 + 3) hrs, for
Job - 2 = 2 hrs. A (/I), B (/2), C (/2), D (/2), E (fl)
Sequencing 503

Flow chart for Sequencing Problems :

Start

A.
Formulate the information
into sequencing table

Yes
Use graphical
Stop
method to solution

Use the sequencing rule like

Yes FIFO/LIFO/EDD/SPT/LPT/ Stop
SIRO/preferential etc.

Apply Jackson's rules

No Min. process time of midiacre
series > Max. Process time
of first or/and last series

Yes

Apply Johnson's Rule

- If Min. is in first series place,
it left available Revise to convert into
- If Min. is second series place, 2M/Cxn jobs case
it most right available position

Delete allocated position

and job and repeat

Compute time schedule

chart and Gantt chart

FIGURE 9.4 :
504 Operations Research : Theory and Practice

Review Questions
1. What are the assumptions made in sequencing problems? [JNTU (Mech.) 98/S]

2- Describe the procedure of optimal solution for processing n jobs through two
machines. [JNTU (Mech.) 98/S]

3. Describe the procedure of optimal solution for processing n jobs through three
machines and hence for n jobs through m machines.
[JNTU (CSE) 96]

4. Explain the rules of sequencing V jobs when there is only one machine to
process.

5. Describe various sequencing models. [JNTU (Mech.) 97/P]

6. Explain the working rules of n jobs X 2 machines and n jobs X 3 machines with
suitable examples [JNTU (Mech.) 98/P]

7. Explain the situations leading to multiple optimal solutions in sequencing

problems. [JNTU (Mech.) 97/CCC]

8. Explain the possibility and working rules of a maximisation case in sequencing.

9. Explain the Jackson's conditions in working with njobsX 3 machine sequencing

problems.

10. Explain the relevance and importance of Gantt chart in sequencing.

11. Write a short note on the contributions of S.M. Johnson & Jackson towards
Operations Research. [JNTU (Mech.) 94]

12. Explain the algorithm for sequencing of 2 jobs to process on V machines.

[JNTU - (Mech.) 95, Mech/Mechatronics/Chem. 2001/S]

13. Explain "No passing rule" with reference to sequencing problems.

[OU (Maths) 85]

14. Give justification of Johnson's rule for sequencing n jobs X 2 machines.

[OU (Maths) 87]

15. Write the sequencing algorithm for 5 jobs to process on 3 machines.

16. Give justification of graphical method used to sequence 2 jobs X M machines.

[JNTU (Mech.) 98/S]
Sequencing 505

17. Explain the elements that characterise the sequencing problem.

18. Give different examples of sequencing problem of various models from your
daily life. [JNTU Mech/Prod - 99, CSE/ECE 2001]

19. Discuss the following terms in the context of sequencing problems.

(a) No. of machines (b) Waiting time ofjob
(c) Idle time of machine (d) No passing rule

20. Critically appreciate the significance of job sequencing in a job/batch

production. How this importance would vary in mass/flow production.

Objective Type Questions

1. The time required for printing of four project reports P,Q,R and S is 4,7,9 and
6 hrs while its data entry requires 6,3,2 and 5 hrs respectively. The sequence
that minimises the total elapse is
(a)PSQR (b)PQRS (c) PSRQ (d) RQSP

2. Printing of five books P,Q,R,S and T takes 9,3,4,6 and 7 hrs while for binding
they take 6,5,8,5 and 2 hrs respectively. The correct sequence that maximises
the total elapse is

(a)QRPST (b)TSPRQ (c) PTSQR (d) RQSTP

3. In which of the following we get multiple optimal solutions
(a) Same least time for two jobs in both series of machines of %m/c x n jobs
sequence model.
(b) If we can connect the zeros as the corners of a rectangular loop, allocated
cells alternatively in AP or TP
(c) If objective function has same slope to that of one of non-redundant
constraint of an LPP.
(d) all the above

4. Which of the following is wrong in assumptions of sequencing 2 machines x n

jobs by Johnson's Rule

(a) no passing is allowed

(b) processing times are known
(c) time of moving job from one machine to the other is negligible
(d) the time of processing is dependent on order of processing
506 Operations Research : Theory and Practice

5. According to the assumptions of 2m/c x n job sequencing

(a) once a job loaded can be removed before completing

(b) two jobs can be simultaneously processed on a machine
(c) machines used are different type
(d) the order of completion ofjobs are dependent on sequence ofjobs

6. According to Johnson's rule, the smallest processing time if occurs in

(a) first series, place it in the first available position

(b) second series, place it in first available position
(c) first series, place it middle
(d) second series, place it middle.

7. The sequencing of 2>m/c x n jobs should pass Jackson's rule according to which
Maximum process time of2nc* series

(a) > minimum process time of lstor3 series

(b) > maximum process time of lstor3rd series
(c) < minimum process time of lstor3 series
(d) < maximum process time of \stor2> series.

8. Which of the following does not characterise the sequencing problem

(a) no passing rule
(b) idle time of machine
(c) number of machines and number ofjobs
(d) number of movements while processing the jobs.

9. FCFS rule is most applicable to case of sequencing

(a) lm/c x n jobs (b) 2m/c x n jobs

(c) Sm/c x n jobs (d) nm/c x 2 jobs

10. The sequencing rule usually followed at a petrol bank when 'n' vehicle are
waiting, is
(a) FIFO (b) LIFO
(c) lowest process time (d) Highest profit rule
Sequencing 507

11. If two jobs/! and/ 2 have same minimum process time in Mj series but process
time/! is less than that of/2 in M2 series then/! occupies

(a) first available place from left

(b) second available place from left
(c) first available place from right
(d) second available place from right

12. If two jobs Ji and / 2 are in tie at M 2 series but for Aflfprocess time for/j is
less than that of/2, then among the available positions,/ 2 occupies

(a) last position (b) last but one position

(c) first position (d) second position

13. If two jobs/! and/ 2 have same process times in both series we prefer

(a)/i (b)/ 2
(c)/ 2 and/ 2 (d)/ } or/ 2

14. Multiple optimal solutions are obtained if two jobs have same process times in

(a) first m/c series (b) second m/c series

(c) both series (d) any one series

15. If a job is has minimum process time in first m/c series as well as in second m/c
series, the job is placed

(a) available first position (b) available last position

(c) any one (first or last) position (d) both first and last position.

16. jobs A to E have process times as 12,4,20,14 and 22 on Ist M/C and 6,14, 16,
18 and 10 on 2™*M/C, the optimal sequence is
(a) CEABD (b) AECDB
(c) BCDEA (d) BDCEA
17. In the above problem, if machines have idle times 6 and 14 hrs on Mi &M 2
respectively. Then total elapse = hrs.
(a) 66 (b) 78
(c) 58 (d) 86
508 Operations Research : Theory and Practice

18. Consider the following process times (in hrs)

Jobs R
5 1 9 3 10

M2 2 6 7 8 4

The optimal sequence is

(a)PTRSQ (b)QSRTP (c) QRSTP (d) PTSRQ

19. In the above problem idle times for My andM 2 are 2 and 3 hrs respectively.
The total elapsed time = hrs

(a) 28 (b)33 (c)30 (d)31

20. In 3 m/c x 5 jobs case, the least of processing times on machines/I, B and C are
5, 1 and 3 while highest of process times are 9, 5 and 7 respectively. The
Jackson's rule is applicable if order of the machines is

(a)B-A-C {b)A-B-C
(c)C-B-A (d) any order

21. In a maximisation case of sequencing 2M/C x n jobs the job is placed at available
left first position if it has process time in machine series

(a) least, first (b) highest, first

(c) least, second (d) highest second

22. The fundamental assumption for Johnson's method of sequencing is

(a) no passing rule

(b) passing, rule
(c) similar machines must be used for first and second process
(d) non zero process time

23. If a job has zero process time for any machine the job must

(a) posses first position only

(b) possess last position only
(c) possess extreme position
(d) be deleted from the sequencing
Sequencing 509

24. No passing rule means

(a) a job once loaded on a m/c should not be removed until it is completed
(b) a job can not be processed on second m/c unless it is processed on first machine
(c) a machine should not be started unless the other is ready to start
(d) no job should be processed unless all other are kept ready to start.
25. A sequencing problem may be infeasible or non optimal or not applicable in
case of

(a) 1 M/C x n jobs (b) 2 M/C x n jobs

(c) 3 M/C x n jobs (d) 2 jobs x n M/C

26. If passing is allowed, the problem can be solved by

(a) Johnson's (2 M/C x n jobs) rule

(b) Jackson's (3 M/C x n jobs) rule
(c) hungarian floods technique
(d) none of the above
27. The technological order of machines to be operated is fixed in

(a) 1 M/C x n jobs (b) 2 M/C x n jobs

(c) 3 M/C x n jobs (d) n M/c x jobs

28. in a 2 jobs x n M/C sequencing, a line at 45° represents

(a) job 1 is idle (b) job 2 is idle

(c) both jobs are idle (d) no job is idle

29. In 2 jobs x n machine case, the elapsed time for job 1 (on X axis ) is reported
as

(a) process time ofjob 1 + idle time ofjob 2

(b) process time ofjob 2 + idle time ofjob 1
(c) process time ofjob 1 + total vertical distance
(d) process time ofjob 1 + total horizontal distance
510 Operations Research : Theory and Practice
30. Given process time for 6 jobs on three machines P, Q and R as follows.

Job 1 1 2 3 4 5 6
M/CP 12 10 9 14 7 9
M/CQ 7 6 6 5 4 4
M/CR 6 5 6 4 2 4

The optimal sequence can obtained if the order of M/C to process is

(a) PQR (b) PRQ

(c) QPR (d) RPQ

Fill in the Blanks

1. If there is some least process time for two jobs in both series, then the
sequencing problem will have

2. The basic assumption for Johnson's rule for 2 M/C x n jobs sequencing is that
is not allowed

3. According to Johnson's rule if smallest processing time appears in 2nd M/C

series, then its position in the sequence is

4. The sequence of 3 M/C x njobs is converted to 2 M/C x njobs if

5. FIFO, LIFO, SPT, LPT etc. are the rules of sequencing

6. If in two jobs, A has same process time as that of B in 1st machine series but
greater in 2nd machine series, then the most left available position will be
occupied by

7. The most min. process time for {Jy , M2) is equal to that of (/2, M2) and process
times for (Jlf Mx) greater than that of (/2, M{) in a 2 M/C x 5 jobs sequencing.
Then position of/2 is

8. The position of/^ in the above question is

9. If/j and/2 n a v e equal process times in bothMj and M 2 series, then we choose
Sequencing 511

10. If min. process time appears in both My series andM 2 series for the same job,
then the job is placed of available positions

11. Jobs P,QR and S have process times 10, 11,8 and 6 on machine Mj and 9, 7,
5, 6 on machine M2 then the sequence is or

12. In a maximisation case of sequencing 2 M/C x n jobs, a job is placed in left (first)
available position if it has process time (or profit) in first M/C
series

13. If process time for job is zero in 1 st series, then this job should occupy
position

14. The idle time of second machine decreases if the job with
process time is taken first on first machine

15. No possing means

16. If passing is allowed, the sequencing problem becomes a case in

problem.

17. In 2 jobs x n M/C sequencing is fixed

18. In a 2 jobs x n M/C case, a horizontal line represents the idle time of
assuming/! and/ 2 an X and Y axes respectively.

19. In the graphical solution for 2 jobs x n machines, when both jobs are processed,
then line moves at

20. For bindingfivebooks P, Q R, S and T take 9, 3,4, 6 an 7 hrs while for printing
they take 6, 5, 8, 5 and 12 hrs respectively. The optimal sequence is

Answers
Objective Type Questions :
l.(d) 2.(c) 3.(d) 4. (d) 5.(c)
6. (a) 7.(c) 8.(d) 9. (a) 10. (a)
11. (b) 12. (a) 13. (d) 14. (c) 15. (c)
16. (d) 17. (b) 18. (b) 19. (c) 20. (b)
21. (b) 22. (a) 23. (c) 24. (b) 25. (c)
26. (c) 27. (d) 28. (d) 29. (c) 30. (b)
512 Operations Research : Theory and Practice

Fill in the Blanks :

1. Alternate or multiple optimal solution 2. Passing
3. least available or extreme right 4. min. process time of 1st or/and 3rd M/C
series > max. process times of 2nd M/C

5. 1 M/C xn jobs 6. job B

7. fourth 8. fifth
9. arbitrarily/either/i or/2 10. first or least/left or right

ll.SPQRorPQSR 12. largest or highest

13. first position 14. lowest or minimum
15. A job can not be processed on 2nd M/C 16. assignment
until it is processed on 1st machine
17. Technological order of M/C 18. Job 2

19. 45° 20. SPTRQ

Replacement Analysis 515

10.0 Introduction
No machine is immortal and immune completely to any failures. No matter
how safely you run, how closely you follow the instructions of the manufacturer or
supplier, how best you maintain to its standards and specification. Perhaps we can
only try to prevent or prolong the occurrence of failure if we know the probable reason
for its occurrence. Even in such cases, some times we can temporarily stop the
occurrence of the failure and some other times we can reduce its impact or volume of
failure. However, this is possible if we have the complete knowledge of the failures
that may occur on the given equipment and their causes, effects or costs and the
remedial measures. The awareness of equipment failures often makes the engineer
so confident that after the rectification of the failure he will be able to assure the
production manager about its running condition.

10.1 Failure Patterns

1011 Definition
Failure is defined in many ways. Of them, the popular and appropriate are
given below.
Failure is defined as inability of a machine or equipment to perform the
intended or specified job under specified conditions.
A failure is defined as an event that changes a product or machine or equipment
from an operational condition to a non-operational condition.
Failure can be defined as "Non-conformance to some defined performance
criterion".

10.1.2 Behaviour of Machines

Almost all machines or equipment whether electrical or mechanical or
electronic, assumed to behave in the same manner. These machines are normally
expected to have one of the three types of behaviours with reference to the failures
that occur. These three behaviours are as follows :
1. The rate of failures is decreasing.
2. Trie rate of failures is constant.
3. The rate of failures is increasing.
Amazingly, every machine is found to have all the three behaviours
siginificantly and distinctly in certain period of its lifetime. Therefore, it has gained
so much of significance in replacement analysis studies.
Machine Life Cycle (MLC) - Bath Tub Curve :
Machine Life Cycle is classified into in three phases, which is analogous to the
three phases of Human Life Cycle as shown below.
L Infancy Phase : Early failures or infant failures.
2. Youth Phase : Random failures or rare event failures.
516 Operations Research : Theory and Practice

3. Old Age Phase : Wear out failures or old age failures

Increasing failure
Decreasing failure Random failure rate in old age
rate in infant stage rate in youth stage
Start up Chance Wear out
failure — failure — failure Reduction of
period period period failures through
maintenance

Time
| Suitable maintenance policies for various stages in machine life cycle |

Contaractual maintenance Preventive maintenance, Replacement, recondition

or reffering to original breakdown/corrective operate to failure &
equipment manufacturer maintenance/condition corrective maintenance etc
due to guarantee based/time based/failure or some condition based
/warrantee based maintenance maintenance techniques
techniques

FIGURE I O.I : BATH TUB CURVE

1. Early Failures or Failures at Infant Stage of Machine :

The machine immediately after it is brought from the Original Equipment
Manufacturer (OEM), may not work with full efficiency due to various reasons such
as the initial friction between the moving or rotating parts, not adjusting to the
environment, lack of skill and knowledge of the person operating on it etc. To control
these failures one has to know about the machine thoroughly and follow the
instructions strictly. Inspite of following the instructions given by OEM strictly, the
outptput may be low or slow or delayed. Yet, it is necessary in the interest of good running
andi Llong life of the machine. For example, when you buy a scooter, the OEM gives
you instructions that you have to run the vehicle at not more than 40 KMPH up to first
1500 km.This obviously restricts and slows down the job and may cause the
inconvenience to the user. But if this is not followed it may lead to a catastrophic failure
by affecting the piston movement in the cylinder or high fuel consumption which will
be uneconomical. Such problems are very common to any new machine. However,
the period of these infant failures may vary from one machine to the other. (See figure
and table). During this failure period, in the event of machine is usually referred to
OEM (of the machine is covered under guarantee.warantee) or any contractual
maintenance.
2. Random or Rare - Event Failures :
These failures occur in young stage of the machine. After passing over the
infant stage, the machine will be running with its full efficiency and the user will enjoy
its full fruit in this period only. The effective usage and correct maintenance can
enhance this portion of its life. User's care may come down in this stage due to the
facts that the user might have got boredom or monotony or some sort of negligence
as he observes the machine will be working with full efficiency though much care is
not taken. This act in fact may or may not result as failure immediately, but its impact
will be there on long run by affecting its life and wear out of the parts etc. However,
the immediate failures are known as random or rare-event failures. The reliability of
machine will be very high in this stage. Usually preventive maintenance and
breakdown maintenance are adapted in this stage to low cost failures wliile RCM
(Reliability Centered Maintenance) and CBM (Condition Based Maintenance)
techniques are employed on the machines that are high cost or in case the failure may
lead to high cost or damage. (See figure and table)
Replacement Analysis 517
3. Old Age of Wear Out Failures :

Most of the efforts put by maintenance are attributed to this stage and costs
could go high at this stage if this stage is not detected. Calender time is not only the
scale for detection of this stage. For instance a machine used sparingly and a machine
used continuously will not reach to the old age stage at the same time, though they
are bought at same time. Hence one should notice that the operational period,
conditions maintained while usage, care and efforts put on it to increase its life during
its infant and young stages etc., are a few factors governing the old age failures wever,
these failures are mainly because of the worn out parts of the machines. As and when
this stage is noticed in the machine, a plant engineer may have to choose one of the
following alternative strategies. (See table and figure)
• Replacement of the machine with the new one.
• Reconditioning of the machine.
• Updating with the new technological features.
• Operate to failure and corrective maintenance (as long as its average annual
maintenance cost is less than or equal to the interest on the cost of the new
machine) or selling in seconds.
• Scrapping.

It is important to note here that the age of a machine (infant/young/middle/old) is

not just decided by the calendar time but by its running time and other factors such
as environment, usage etc.

The above phases are summarised in the following tabular form.

TABLE:
Phase Type of Failure Cause of Cost of Suitable
Failure Rate Failure Failures Maintenance Policy
I n f a n t Early failures Decreasing Faulty Design, Erratic Medium (Under warantee /
Phase or infant operation, Environment to high guarantee) refer to OEM
failures problems, Installation
errors, Initial friction Contractual maintenance
(No warantee)
Y o u t h Random or Constant Operation errors, Heavy Low to Break-down maintenance,
Phase Chance or workloads, Over run medium Preventive maintenance
Rare-Event
failures Medium Reliability centered
to high maintenance, condition
based maintenance
Olde Age Wear-out or Increasing Wear, Tear, Creep, Fatigue, Low Operate to fail and
Phase Age failures Weakened parts corrective maintenance
High Reconditioning or
replacement policies or
scrapping
518 Operations Research : Theory and Practice

10.13 Failures Based on the Volume of Failure

1. Small Failures : These are the failures, which can be rectified in few minutes.
These failuers are very common on any machine. These will not have any
considerable impact on the machine performance and on the operator.

2. Minor Failuers : These are the failures, which can be rectified in a few hours
or with a little effort. The failures of this kind will have a little to considerable
effect on the productive work and could hardly damage the machine or men
or environment

3. Major Failures : The^e failures take few days to rectify. They also may require
large work force, knowledge and skill to rectify. These failures may cause
considerable damage to the machine, minor injuries to men and affects the
regular work,

4. Catastrophic Failures : These are the costliest failures. The occurrence of such
failures may cause the damage to the machine, men and some times the
environment also. The rectification or recovery from the losses of this failure
may take a few weeks to months even.

10,1*4 Types ofFailures Based on the Mode of Failure

1. Sudden Failures :These types of failures occur in items after giving some
period of desired service rather than deterioration while in service. This period
of giving desired service is not constant but follows some frequency distribution,
which may be progressive, retrogressive or random in nature.

(a) Progressive Failure : If the probability of failure in the beginning of an

item is less and gradually increases in its life, then such failure is called
progressive failure. For example, light bulbs and tubes fail progressively.

Time — Time • Time •

(a) Progressive Failure (b) Retrogressive Failure (c) Random Failure Pattern

FIGURE 10.2 ; FAILURE PATTERNS OF SUDDEN FAILURES

Replacement Analysis 519

(b) Retrogressive Failure : If probability of failure in the beginning of the life

of an item is more but as time passes the chances of its failure become less
then such failure is said to be retrogressive.

(c) Random Failure: In this type of failure, the constant probability of failure
is associated with items that fail from random causes such as physical
shocks, not related to age. For example, vacuum tubes in air-burn
equipment have been found to fail at a rate of the age of tube.

2. Gradual Failures : Gradual failure is progressive in nature, i.e. as the life

increases, its operational efficiency also deteriorates resulting in increased
running (maintenance and operating) costs. They also cause decrease in its
productivity and decrease in the resale or salvage value. Mechanical items like
pistons, rings, bearings etc., and automobile tyres fall under this category.

Running time •

FIGURE! 0 . 3 : GRADUAL FAILURE PATTERN

10.2 Failure Costs

Cost Involved in Machine Failure Analysis :

While analysing the machine failure we are concerned with the following costs.
1. Purchase cost of machine or equipment capital investment.
2. Depreciation or salvage value or scrap value.
3. Running costs including Maintenance, Repair and Operating (MRO) costs.
4. Fialure costs and damage costs.

In the above four, the first three are inevitable, but failure costs can be prevented
by better maintenance policies. On critical examination, we can notice that the first
three costs depend on age of the machine. These costs vary with the running age as
follows.
520 Operations Research : Theory and Practice

1. Purchase costs of machine or equipment is considered to be independent of

age of machine. However the interest on the investment is assumed to be lost.

2. Resale value or salvage value or scrap value of the machine decreases with the
running age of the machine. As the machine grows old, its resale value comes
down. This decrease depends on actual or expected condition of the machine.

3. Running cost or operating cost or maintenance cost of the machine is due to

its minor failures or preventive maintenance or operating costs etc. These will
increase on the machine as the age grows. This is assumed to be increasing due
to the wear and tear on the moving parts of the machine.

The above costs are shown graphically on a hypothetical machine here below.

Running cost

Purchase cost of machine

Optimum level
of service

Opt. Replacement age

'Salvage cost
Age of the Machine

FIGURE 10.4 : COST INVOLVED IN REPLACEMENT ANALYSIS

Summing, the above three costs, we can notice that the average total cost
decreases for certain period and then increases. The age when the graph shows its
minimum costs will be optimum age of replacing the machine.
Necessity or Significance of Replacement:
The replacement of parts or entire machine will become significant and
necessary in the following cases.
1. When average costs of repairs or maintenance or operating goes higher than
the costs of machine. Or in other words, the cost of maintenance will increase
to such an extent that the average annual repair or maintenance cost is greater
than or equal to costs of new machine.

2. When the machine completely fails to work very frequently by which the
production schedules are interrupted.

3. Machine runs with less efficiency and therefore not economical.

Replacement Analysis 521
4. Modified or new designs in the market may give an edge of advantage such as
reduced cost of production or ease and comfort of operation or more functions
are available in new design etc.

5. If it is expected that the existing model may become absolete or resale value
may drastically come down.

6. If the item is non-repairable type or use & throw type.

10.3 Types of Replacement Problems

The replacement of machine is considered as the following three types in this
book.
Type - / : Replacement policy for items when money value remains unchanged with time.

Type - II: Replacement policy for items when money value changes with time.

Type - III: Group replacement policy.

These are explained in detail with examples and illustrations in the sections to
follow,

10*4 Type -1: Replacement Policy when Money Value does not
Change with Time
Let us now find the optimal policy for the case of replacement when money
value does not change with time.
Let C = capital or purchase cost of new item.
S = scrap or salvage or resale value of the item at the end of/ years.
R (t) = running cost for the yeat t.
n = replacment age of the equipment.
Here two cases arise.
Case (i) : When time t is a continuous variable :

If the equipment is used for T years, then the total cost incurred over this period
is given by
Tc = capital (or purchase) cost - scrap value at the end of t years + running
cost for t years.
522 Operations Research : Theory and Practice

Theretore the average cost per unit time incurred over the period of n years is

ATCn =-\C- . . . . (1)

To obtain optimal value of n for which ATCn is minimum, differentiate ATCn

with respect to n and set the first derivative equal to zero. i.e. minimum of ATCn.

or R(n) = -\C-S+\R(f)dt

R(n) =ATCn • • • (2)

Hence the following replacement policy can be derived with the help of
equation (2)

Policy : Replace the equipment when the average annual cost for n years becomes equal to the
current/annual running cost.

i.e. R(n) = - \ c - S + \R(t) dt

Case (ii) : When time t is a descrete variable :

The average cost incurred over the period n is given by

n
ATCn=±\C-> .... (3)

are
If C-S and ^ ^(0 assumed to be monotonically decreasing and increasing

respectively, then there will exist a value of n for which ATCn is minimum. Thus we
shall have inequalities

ATCn_1-ATCn>0
and ATCn+l-ATCn>0
Replacement Analysis 523

Rewriting equation (3) for period n + 1, we get

^R(t)+R(n + 1)
n+l

C-S+
I (n+l)
n+l n+l

n+l n+l

Therefore ATCn+ x -ATCn = ATCn + * -ATC,

n+l

n+l | n+ 1- l l
1)
n+l n+l

Since ATCn+ x ~ATCn > 0 we get

Rin+l)_ATC2L>Q
(n+l) n+l

R(n+l)~ATCn>0

R(n+l)>ATCn . . (4)

Similarly ATCn _ x -ATCn > 0 implies that R (n) <ATCn^x.

This provides the following replacement policy.

Policy 1 : If the next year, running cost, R (n + I) is more than average cost of nth year, ATCn
then it is economical to replace at the end ofn years

i.e. R(n+l)>~\c-S + TR(t)

n ^
t=0
524 Operations Research : Theory and Practice
Policy 2 : If the present year's running cost is less than the previous year's average cost, ATCn _ j
then do not replace.

i.e.

The Procedure:
<=°
Thus the procedure for obtaining the decision when to replace the equipment
in the case of money value not changing with time can be outlined as follows.

Step 1: Draw the table with columns as shown below and enter the values of column 1,
2, 3 and 4 as given in the problem.
Cumulative Average total cost
Year of Cost of Salvage Running Total Cost (7T) =
Cost Net value running cost = ATC =
service equipment value
(C-S)
(«) (C) (5) Rn
I*.
(C~S) + %Rn ^(c-s) + S^l
(1) (2) (3) (4) (5) = (2) - (3) (6) = £ (4) (7) = (5)+ (6) 8 = (7)/(l)

Step 2 : Calculate net value by difference of cost and salvage value (C-S) i.e., 2nd column
value minus 3rd column value and enter in column 5

Step 3 : Calculate cumulative running cost i.e., ^T Rn cumulation of column 4 and enter
in column 6.

Step 4 : Calculate total cost TC = (C-S) + £ Rn i.e., sum of 5th column value and 6th
column value to enter in col. 7.

Step 5: Calculate average total cost for n years, i.e., divide 7th column value by 1st
column value i.e., ATC = — UC-S) + ]£ RJ and enter in col. 8.

Step 6: Observe the values in column 8 and identify the minimum value. The year
corresponding to this minimum value is the age of the equipment to be
replaced.
Replacement Analysis 525

This is illustrated through the numerical example given below.

ILLUSTRATION 1
A firm is thinking of replacing a particular machine whose cost price is Rs.
12,200. The scrap price ofthe machine is only Rs. 200. The maintenance costs are
found to be asfollows:
Year 1 2 3 4 5 6 7 8
Maintenance cost 220 500 800 1200 1800 2500 3200 4000
Determine when thefirm should get the machine replaced. [JNTU B. Tech. (CSE) 97/S]
Solution i
The calculations of average running cost per year during the life of the machine
are shown in the following table :
Cost price machine (C) = 12,200 ; 5 = 200; C~S == 12,000
Year Cost Salvage Running Depreciation Commulative Total Cost Average
n C value (Rs) cost (Rs) Cost (Rs) Running Cost (Rs) (Rs) Cost (Rs)
S R(n) C-5 ^R(n) TC Per Year At
cn
(1) (2) (3) (4) (5) = (2)-(3) (6) = 2 (4) (7)=(5)+(6) (8)=(7)/(l)
1 12200 200 220 12000 220 12220 12220
2 12200 200 500 12000 720 12720 6360
3 12200 200 800 12000 1520 13520 4506.67
4 12200 200 1200 12000 2720 14720 3680
5 12200 200 1800 12000 4520 16520 3304
;?";&&&... •;• A-itf» <- 1
7 12200 200 3200 12000 10220 3174.29 22220
8 12200 200 4000 12000 14220 3277.5 26220
From the above table it may be noted that the average cost per year, ATCn is
minimum in the sixth year (Rs. 3170). And this average cost is increasing from 7 year
onwards. Hence the machine should be replaced after 6 years.
ILLUSTRATION 2
A plant Manager is considering replacement policy for a new machine. He
estimates thefollowing costs in Rupees.
Year 1 2 3 4 5 6
Replacement cost at beginning of year 100 110 125 140 160 190
Salvage value at end of year 60 50 40 25 10 0
Operating Costs 25 30 40 50 65 80

Find an optimal replacementpolicy and corresponding minimum cost.

[JNTU (Mech.) 98/S]
526 Operations Research : Theory and Practice

Solution :
The calculations for replacement of the machine are shown in the following
table :
Year Replacement Resale value Net Value Operating Commulative Total Cost Average
Cost (S) Cost Running Annual cost
Cost 2 Rn ATCn

(1) (2) (3) (4)= (5) (6) = E(5) (7)= (4) +

(2) - (3) (6)

100 60 40 25 25 65 65

125 40 85 40 95 180 60

140 25 115 50 145 260 65

160 10 150 65 210 360 72

190 190 80 290 480 80

From the above table, it may be noticed that the average cost per year, ATCn
is minimum in the 2nd year, i.e., ATC2 is Rs. 57.5 which is less than^TCj (Rs. 65) and
ATC$ (Rs. 130/-). Hence the machine should be replaced at the end of second year.
ILLUSTRATION 3

A fleet ownerflndsfrom his past records that the costper year of running a vehicle
whosepurchase price is Rs. 50000 are as under:

Year 1 2 3 4 5 6 7
Running cost Rs. 5000 6000 7000 9000 21500 16000 18000
Resale Value Rs. 30000 15000 7500 3750 2000 2000 2000

There after running costs increase by Rs. 2000, but resale value remains constant
atRs. 2000. At what age is a replacement due?
[JNTU B.Tech. (Mech.) 97/CCC, (ECE) 97/CCC]

Solution :
The required calculations are shown in the following table :
ATC
Year C S C-S R(n) cum R(n) TC
(1) (2) (3) (4) =(2)-(3) (5) (6) = 2 (5) (7)=(4)+(6)
(1)
1 50,000 30,000 20,000 5000 5000 25,000 25,000

2 50,000 15,000 35,000 6000 11,000 46,000 23,000

Replacement Analysis 527

3 50,000 7500 42,500 7000 18,000 60,500 20166.6

4 50,000 3750 46,250 9000 27,000 73250 18312.5
5 50,000 2000 48,000 21500 48,500 96,500 19300
6 50,000 2000 48,000 16000 64,500 112500 18750
7 50,000 2000 48,000 18000 82,500 130500 18642.8
8 50,000 2000 48,000 20000 102500 150500 18812.5
9 50,000 2000 48,000 22000 124500 170500 18944.4

The machine should be replaced at the end of 4 year or at the end of 7 year
if it is not replaced at the end of 4th year.
ILLUSTRATION 4 —
Machine A cost Rs. 45,000 and the operating costs are estimated ofRs. 1000 for
the first year, increasing by Rs. 10,000 per year in the second and subsequent
years. Machine B costs Rs. 50,000 and operation costs are Rs. 2000 for the first
year, increasing by Rs. 4000 in the second and subsequentyears. If we now have a
machine of type A, should we replace it with B? If so when? Assume that both
machine have no resale value and future costs are not discounted.
[JNTU (Mech.) 95, (CSE) 2000/S]

Solution :

The calculations of average costs running per year during the life of machine
A and B are shown in tables give below.
TABLE: CALCULATIONS OF AVERAGE RUNNING COST FOR MAC HI WE A.
Year of Service Running Cost Cumulative Depreciation Total Cost (Rs.) Average Cost (Rs)
in Rs. R(n) Running Cost in Cost (Rs.) TC Rs. ATCn
Rs. £ R(n) C-S

(1) (2) (3) (4) (5) = (3) + (4) (6) = (5)/(I)

1 1000 1000 45,000 46,000 46,000
2 11,000 12,000 45,000 57,000 28,500

ifiiMi
64,000 45,000 1,09,000 27,200
31,000
1,05,000 45,000 1,50,000 30,000
41,000
1,56,000 45,000 2,01,000 33,500
51,000
From the above table it may be noted that the average running cost per year is
lowest in the third year, i.e., Rs. 26,000. Hence, machined should be replaced after every
three years of service.
528 Operations Research : Theory and Practice

TABLE: CALCULATIONS OF AVERAGE RUNNING COST FOR MACHINE B

Year of service Running cost Cumulative Depreciation Total cost (Rs.) Average cost
n (Rs.)/2(n) running cost cost (Rs.) C-S TC (Rs.)ATCn

(1) (2) (3) (4) (5) - (3) + (4) (6) = (5)/(I)

2,000 2,000 50,000 52,0*00 52,000
6,000 8,000 50,000 58,000 29,000
10,000 18,000 50,000 68,000 22,667
14,000 32,000 50,000 82,000 20,500

22,000 72,000 50,000

IH
1,22,000 20,333
From the above table it may be noted that the average running cost per year is
lowest in the fifth year, i.e., Rs, 20,000, This cost is less than the lowest average running
cost (Rs. 26,000) per year for machine A. Hence machined should be replaced by
machine 25.
Now to find the time of replacement of machined by machine B, the total cost
of machine A in the successive years is computed as follows :
Year 1 2 3 4
Total cost incurred 46,000 57,000-46,000 78,000-57,000 1,09,000-78,000
(Rs) =11,000 =21,000 =31,000

Machined should be replaced by machine B at the time (age) when its running
cost for the next year exceeds the lowest average running cost (Rs. 20,000) per year
of machined.

Calculations show that the running cost (Rs. 21,000) of machined in the third
year is more than lowest in average cost (Rs. 20.000) per of machine B. Hence machine
A should be replaced by machine B after two years.

Practice Problems
1. The cost of a machine is Rs. 6100 and its scrap value is only Rs. 100. The
maintenance costs are found from experience to be
Year 1 2 3 4 5 6 7 8
Maintenance cost Rs. 100 250 400 600 900 1200 1600 2,000

When should the machine be replaced ? [JNTU (Mech.) 93]

Answer: At the end of sixth year Rs. 1583.33
Replacement Analysis 529

2. A truck - owner finds from his past experience that the maintenance costs are
Rs. 200 for the first year and then increase by Rs. 2,000 every year. The cost of
truck type A is Rs. 9,000. Determine the best age at which to replace the truck.
If the optimum replacement is followed what will be the average yearly cost of
owning and operating the truck? Truck type B costs Rs. 20,000. Annual
operating costs are Rs. 400 for the first year and then increase by Rs. 800 every
year. The truck owner has now the truck type A which is one year old. Should
it be replaced by B type, and if so, when?
Answer; After every 3rd year, trucks is to be replaced (Rs. 5200). After every 6th
years truck B (Rs. 5733.3) is replaced. Comparing both, cost of truck B is
lower than that of Truck A from 3rd year onwards, hence after 3rd year
replace Trucks by Truck B.
3. (a) Machined costs Rs. 9000, Annual operating costs are Rs. 200 for the first
year, and then increase by Rs. 2,000 every year. Determine the best age
at which to replace the machine. If the optimum replacement policy is
followed, what will be the average yearly cost of owning and operating the
machine?
(b) Machine B costs Rs. 10,000. Annual operating costs are Rs. 400 for the
first year, and then increase by Rs. 800 every year. You now have a machine
of type A which is one year old. Should you replace it with B, if so, when?
[ICWA, June 1985]

Answer: (a) Total cost in the 3rd year is least i.e., Rs. 5200 hence replace at the end
of 3rd year.
(b) Average annual cost is least at the end of 5th year. The total cost Rs. 2200
for 1 year old machine A remains less than minimum average cost 400/-
on machine B until second year, therefore replace machine A before 3rd
year.
4. A firm is considering replacement of a machine whose cost price is Rs. 12,200
and the scrap value is Rs. 200. The maintenance costs are found from
experience to be as follows :
Year 1 2 3 4 5 6 . 7 8
Maintenance cost (Rs) 200 500 800 1200 1800 2500 3200 4000
[JNTU (CSE) 96/S, (Prod./Mech,Mechatroirics/Chem.) 2000]

Answer : Replace at the end of the 6th year

5. The following table gives the running costs per year and resale price of a certain
equipment whose purchase price is Rs. 5000.
2 7
.00

Year 1 3 4 5 6
Running costs (Rs) 1500 1600 1800 2100 2500 2900 3400 4000
Resale value (Rs) 3500 2500 1700 1200 800 500 500 500

At what year is the replacement due? [JNTU (CSE) 94/S,

ICWA (DCC) 86, 85, C.A (May) 82]
Answer: At the end of the 4th year ATCS (2733) >ATC4 (2700) <ATC5 (2740)
530 Operations Research : Theory and Practice
6. A truck owner finds from his past records that the maintenance costs per year
of a truck whose purchase price is Rs. 8000, are given below :
Year 1 2 3 4 5 6 7 8
Maintenance cost (Rs) 1000 1300 1700 2200 2900 3800 4800 6000
Resale Price (Rs) 4000 2000 1200 600 500 400 400 400

Determine at what time it is profitable to replace the truck.

[JNTU (Mech.) 93]

Answer: Replace at the end of the 5th year

7. A fleet owner finds from his past records that the costs per year of running a
vehicle whose purchase price is Rs. 5000 are as under :

Year 1 2 3 4 5 6 7
Running costs (Rs) 500 600 700 900 2150 1600 1800
Resale value (Rs) 3000 1500 750 375 200 200 200
Thereafter, running cost increases by Rs. 200 but resale value remains constant
at Rs. 200. At what age is a replacement due? [JNTU (CSE) 96]

Answer: At the end of 5 th year

8. A truck owner from his past experience estimated that the maintenance cost
per year of a truck whose purchase price is Rs. 1,50,000 and the resale value of
truck will be as follows.
Year 1 2 3 4 5 6 7 8
Maintenance cost 10,000 50,000 20,000 25,000 30,000 40,000 45,000 50,000
(Rs)
Resale value (Rs) 1,30,000 1,20,00 1,15,000 1,05,000 90,000 75,000 60,000 50,000

Determine at which time it is profitable to replace. [JNTU (CSE) 95]

Answer : Though the minimum value of the average cost occurs at the end of first
year, it is impractical to replace the truck every year, so next time the
minimum average cost occurs at the end of 4th year. Hence it is better to
replace the truck after every 4 years.
Note : Here the condition after every 4 years, policy maintenance cost increases with time violating if in
the 3rd year.

9. A machine costs Rs. 10,000 operating costs are Rs. 50 per year for the five years.
In the sixth and succeeding years operating costs increase by Rs. 100 per year.
Find optimum length of time to hold the machine before replacing it.
[JNTU (Mech.) FDH 95]

Answer: End of 15 years.

Replacement Analysis 531
10. (a) Machined costs Rs. 3600. Annual operating costs are Rs. 40 for first year
and then increases by Rs, 360 every year. Assuming machine A has no
resale values, determine the best replacement age.

(b) Another machine B which is similar to machined costs Rs. 4000. Annual
running costs are Rs. 200 for first year and then increase by Rs. 200 every
year. It has resale value o' Rs. 1500, Rs. 1000 and Rs. 500. If replaced at
the end of first, second and third year respectively. It has no resale value
during fourth year onwards. Which machine would you prefer to
purchase. Further costs are not discounted. [J*rru (CSIT) 98/S]

Answer: Replace machine A by machine B at the end of 4th year.

11. A truck with first cost of Rs. 80,000 has depreciation and service pattern as
shown below.
Year 1 2 3 4 5 6
Depreciation 28,000 20,000 14,000 5,000 4,000 4,000
Service cost 18,000 21,000 2 ,000 29,000 34,000 40,000

Assume no interest charges. Find age for replacement.

[JNTU (Mech.) 98/CCC]

Answer: After six years

12. A machine owner finds from his past records that the costs per year of
maintaining a machine whose purchase price is Rs. 6,000 are as given below.
Year 1 2 3 4 5 6 7 8
Maintenance cost (Rs) 1000 1200 1400 1800 2300 2800 3400 4000
Resale price 3000 1500 750 375 200 200 200 200
Determine at what age is a replacement due?
[JNTU (Mech.) 97/CCC, (ECE) 97/C, (CSIT) 2000]

Answer : The machine should be replaced at the end of fifth year.

13. The machine owner has three small machines of purchase price Rs. 6000 each
and cost per year of maintaining each machine is same. Two of these machines
are two years old and the third is one year old. He is considering a new machine
of purchase price Rs. 8000 with 50% more capacity than one of two old ones.
The estimates of maintaining cost and resale price for new machine are as given
below ;

Year 1 9 3 4 5 6 7 8
Maintenance cost (Rs) 1200 1500 1800 2400 3100 4000 5000 6100
Resale price (Rs) 4000 2000 1000 500 300 300 300 300
532 Operations Research : Theory and Practice

Assuming that the loss of flexibility due to fewer machines is of no importance,

and he continues to have sufficient work for three of the old machines. What
should his policy be? [JNTU (CSE) 94]

Answer: All the three small machines should be replaced after two years.

16. Fleet cars have increased their costs as they continue in service due to increased
direct operating costs (gas and oil) and increased maintenance (repairs, tyres,
batteries etc.) T h e initial cost is Rs. 3,500 and the trade-in value drops as time
passes until it reaches a constant value of Rs. 500.

Given the cost of operating, maintaining and the trade-in value, determine the
proper length of service before cars should be repalced.

Year of service 1 2 3 4 5
Year end trade-in value 1900 1050 600 500 500
Annual operating cost 1500 1800 2100 2400 2700
Annual maintaining cost 300 400 600 800 1000

[JNTU (Hech. & ECE) 98/S/CCC]

Answer : T h e cars should be replaced at the end of the third year.

10.5 Replacement Policy for Items When Money Value Changes

With Time
Money value changes with time. Suppose you keep Rs. 100 in a bank which
gives you an interest at the rate of 10% p.a. After one year you will recieve Rs. 110.
Thus we can say that "if we have Rs. 100 today, it is worth having Rs. 110 after one
year". Similarly, a rupee possessed today is worth 1.1 after one year @10%.
In the other way, if Rs. 1.10 after one year is Re. 1.00 today, Re. 1.00 after one
year will be 1/1.1 or (l.l)" 1 today.

In a similar fashion one rupee today will be (1.1) two years hence because it is
1.1 after one year and 1.1 x 1.1 in the next i.e., second year. Thus a rupee two years
hence is worth (l.l)" 2 today. Thus the present worth of one rupee, n years hence will
be (1.1 )~n or (1 + 0.10)" today, when money value is changing @ 10%.
Now if the money value is supposed to be changing at the rate of r percent, the
1
present worth of one rupee, n years hence will be (1 + r)^" or
(1+rf
Replacement Analysis 533

Present Worth Factor (PWF) : It is the factor that converts a rupee 'n years hence
changing (money grows but value decreases) at the rate of Y percent is worth today. This is
denoted by v = (1 +r)~ ? \ This is also called "discount rate" or "depreciation rate" or
"present value factor".

Now let C = Initial cost of the equipment

and if R^ = Operating cost in first year, (present worth is also R±)
if 7?2 = Operating cost in second year present worth of
R2 = (l +r)~l R2orv R2
if R§ = Operating cost in third year, and present worth of
/?3 = (1 + r)""2 /? 3 or v2 /?3 and so on

if Rn = Operating cost in nth year present worth of

Thus the present value of all future discounted cost in V year assuming scrap
value of the equipment be zero, is
Pn = C + J?j + v R2 + v 2 £3 + + vn " l Rn

Thus Pn is the amount of money required to pay all future costs of purchasing
the equipment and operating it assuming that it is to be replaced after V years.
Now if we assume that the manufacturer invests the amount Pn by borrowing
at the rate of interest V and repays it in V years with a fixed annual instalment V on
diminishing balance,

Then Pn =.*•+ vx + v 2 1

1-vn
1- V
" 1- v
1-vn

Hence the best period to replace the machine is the period n which minimises
p
1-v n
. Pn. But 1 - v = A positive constant quantity and so we can write Fn = -

and find out the value of n the period at which to replace the machine that minimises
Fn. Since n can assume only discrete values [1, 2, 3, . . . ] we can use the method of
finite differences to calculate its optimal values.
534 Operations Research : Theory and Practice

Fn will be minimum if A Fn _ \ < 0 < A Fn . . . . (b)

Now .... (c)

l-vn

Wehave Pn+l = R2 +.. .+vn vn Rn

Hence

(l-vn+1)(l-
vn(l-v) 1-v"
• Rn + I (e)

1- V »
= A positive constant x
1-v 'Rn+:

Hence Fn has always the same sign as

1-v
From equation (b), n will be optimal if

1- (f)
1-v 1-v
1-VB
From above equation we have — Rn + j - P n > 0

1-v
or or Rn+X>-
n
\ l-vn
C + i?j + v R2 + v- i? 3 + . . . + vn l
Rn
or An+\ 2 n
1+ V +V + . . .+V

r=l
(g)
- 1

r= 1

The right hand expression is the weighted average (denoted by wr) of all costs
up to and including period (n - 1). The weights 1, v, v 2 . . \n~ l
are the discount
factors applied to the costs for each period.
Replacement Analysis 535

From equation (f), the left hand side of expression can be expressed as

r=\
Rn+l< — (h)
r-1

From equation (g) and (h) we conclude that

1. The machine should be replaced if the next period cost is greater than the
weighted average of previous costs.
2. The machines should not be replaced if the next period's cost is less than the
weighted average of previous costs.
Note : When money value is not considered, v = 1
Pn Pn
Rn or Rn . i > — where —
l + 1 + ...tora terms
average yearly cost with no resale value.
This is identical to the previous case when money value was ignored. In real
practice replacement policy is greatly influenced by complicated tax laws prevailing.
Discussion in this regard is not included in the scope of this book. In actual dealings
the influence of tax has got to be taken into consideration.
Steps to Find the Policy when Money Value Changes with Time :

Step 1 : Note the values of capital cost of machine, salvage value, rate of depreciation
PWF etc.

Step 2 : Construct the tabular form as given below, and enter the first 2 columns as per
the given data.
Year Running Cost PWF C + S^v""1 Iv-1 Average Annual
v«-l Total CostT Cn

(1) (2) (3) (4) = (2)-(3) (5) = C+I (4) (6) = I (3) (7) = (5)/(6)

Step 3 : Calculate the present worth factor for each year by the formula — and

put in col. 3.

Step 4 : Calculate Rn vn l
i.e., the product of 2nd and 3rd col. values, and enter them
in col. 4.
536 Operations Research : Theory and Practice
Step 5: Col. 5 is summation of cummulative running cost and capital cost including
(salvage if any) i.e., C + Z Rn vn " l.

Step 6 : Find cummulation of col. 3. and enter in col. 6 [~5 \n ~ *1

Step 7: Calculate the value of col. (5) divided by col. (6) i.e., (c + X Rn vn~l^Z vn~ l

and enter in col. 7.

Step 8 : The lowest value in col. 7. corresponds to the year in which the machine is to
be replaced.
Note : Students are advised to check the problem thoroughly and note whether depreciation is given or
depreciated amount is given or discount rate is given etc.

ILLUSTRATION 5
The initialprice of equipment is Rs. 5000. The running cost varies as shown below.
Year 1 2 3 4 5 6 7
Running cost (Rs.) 400 500 700 1000 1300 1700 2100
Taking a discount rate ofO. 09. Find out the optimum replacement interval.
So luff on :
Year = n Maintenance Discount Discount cost n n ATC
cost = Rr factor = vr ~ * i^v7"1
r=\

(1) (2) (3) (4) = (2)-(3) <5) = C + « 4 , (6) = 1 (3)

(7)=
(6^
1 400 1.000 400 5400 1.000 5400
2 500 0.900 450 5850 1.900 3079
3 700 0,810 567 6417 2.710 2368
4 1000 0.729 729 7146 3.439 2083
5 1300 0.656 853 7999 4.095 1953
MRS
2100 0,531 1115 10117
Hi 1936
7 5.225
From the above table, we observe that ATC$ >ATCe < ATCy
i.e., 1953 > 1921 < 1936 the optimum replacement period is 6 years.
ILLUSTRATION 6
A manufacturer is offered 2 machines A and B* A is priced Rs. 5000 and running
costs are estimated at Rs. 800 for each of the first five years increasing by Rs. 200
per year in the sixth and subsequentyears. Machine B which has the same capacity
as A with Rs, 2500 but will have running cost ofRs. 1200 per year for 6 years,
increasing by Rs. 200 per year thereafter. If money is worth 10% per year, which
machine should bepurchased?
(Assume that the machines will eventually be sold for scrap at a negligibleprice)
[JNTU Mech/Mechatronics/Chem/Prod. 2001/S]
Replacement Analysis 537

Solution \

TABLE FOR A

r=10% (^ = 5000

Year n n 5
Rr
6
r=\

(1) (2) (3) (4)-(2)x(l) (5) = C + Z (4) (6)-I (3) (7) = (5) (6)
1 800 1.0000 800 5800 1.0000 5800.00
2 800 0.9091 727 6527 1.9091 3418.88
3 800 0.8264 661 7188 2.7355 2627.67
4 800 0.7513 601 7789 3.4868 2233.85
5 800 0.6830 546 8335 4.1698 1998.89
6 1000 0.6209 621 8956 4.7907 1869.45
7 1200 0.5645 677 9633 5.3552 1798.81
8 1400 0.5132 718 10351 5.8684 1763.85

Iiiiii
10 1800 0.4241 763 11860 6.7590 1754.70

iTABLE FORB

R=l( CB = 2500
1 1200 1.0000 1200.00 3700.00 1.0000 3700.00

2 1200 0.9091 1090.91 4790.91 1.9091 2509.51

3 1200 0.8264 991.98 5782.59 2.7353 2113.91
4 1200 0.7513 901.56 6684.15 3.4868 1916.99
5 1200 0.6830 819.60 7503.75 4.1698 1799.55
6 1200 0.6209 745.08 8248.83 4.7909 1721.84
1400 0.5645 790.30 9039.13 5.3552 1687.92

§§tI! Ifiitti | | li
9 1800 0.4665 839.70 10699.95 6.3349 1689.25
10 2000 0.4241 848.20 11548.15 6.7590 1708.56

From table for A

ATC& (1763.85) >ATCg (1751.7?) <ATCl0 (1754.70)

Replacement period for^ = 9 year.

538 Operations Research : Theory and Practice

From table for B

ATC7 (1687.9?) >ATCS (1680.23) <ATCQ (1689.25)

Replacement period for 5 = 8 year.

1 -v
The fixed annual payment for - A = xA = •

1-0.9091
xllO97 = Rs. 1752
l-(0.9091)9
1-0.9091

Hence purchase machine B instead of A.

Alternately weighted average cost in 9 year forvl = 1781.72 and that for B in
8 year = 1680.23 which is the lowest.

Purchase B.

Practice Problems
1. A company has the option to buy one of the mini computer : MINICOMP and
CHIPCOMP. MINICOMP costs Rs. 5 lakh and running and maintenance costs
are Rs. 60,000 for each of the first five years increasing by Rs. 20,000 per year
in the sixth and subsequent year. CHIPCOMP has the same capacity as
MINICOMP, but costs only Rs. 2,50,000. However its running and
maintenance costs are Rs. 1,20,000 per year for first five year and increase by
20000 per year thereafter. If the money is worth 10% per years, which computer
should be purchased? What are the optimal replacement periods for each of
the computer? Assume that there is no salvage value for either computer.
Explain your analysis.
[JNTU (CSE) 96/S]

Answer: Better replacement period for CHIPCOMP computer is after six years.

2. Assume that the present value of one rupee to be spent in a years time is Rs.
0.90 and C = Rs. 3000 capital cost of equipment and the running cost are given
in the table as follows.
Year 1 2 3 4 5 6 7
Running cost (Rs) 500 600 800 1000 1300 1600 2000

When should the machine be replaced?

Answer: After 5 years
Replacement Analysis 539

3. An Engineering company is offered two types of material handling equipments

A and B.Ais priced at Rs. 60,000 including costs of installation, operation and
maintenance are estimated to be Rs. 10000. for each of the first 5 years,
increasing every year by Rs. 3000 per year in the sixth and subsequent year.
Equipment B with rated capacity same as A, requires an initial investment of
Rs. 30000 but in term of operation and maintenance costs are estimated to be
Rs. 13,000 per year for the first six years, increasing every year by Rs. 4000 per
year from the seventh year onwards. The company expects a return of 10% on
all its investments. Neglecting the scrap value of the equipment at the end of
its economic life, determine which equipment the company should buy.
[I.I.T.E Grad. 1981, JNTU (Hech. & FDH) 95]

Answer: Purchase J5; Replace after 7 years

4. A manufacturer is offered two machine A and B. A is priced at Rs. 50000 and

running costs are estimated at Rs. 8000 for each of thefirstfiveyears increasing
by Rs. 2000 per year in the sixth and subsequent year. Machine B which has
the same capacity as/I, costs Rs. 25000 but will have running costs of Rs. 12000
per year for six years, increasing by Rs. 2000 per year thereafter.
If money is worth 10 percent per year, which machine should be purchased?
(Assume that the machines will eventually be sold for scrap at a negligible price.
[JNTU (Mech. & FDH) 96]

Answer: Machine B; Replace after 8 year]

5. A truck is priced for Rs. 60000 and running costs are estimated at Rs. 6000 for
each of first 4 years increasing by Rs. 2000 per year in the fifth and subsequent
years. If money is worth 10%, when should the truck be replaced. Assume scrap
value = 0. [JNTU (Mech.) 96/C/S, (ECE) 96/S/CCC]

Answer : Replace after 9 year

6. A machine costs Rs. 10,000 and operating costs are Rs. 500 per year for first
five years. Operation cost increases by Rs. 100 per year in the sixth and
subsequent year. Assuming 10% discount rate of money per year find optimum
length of time to hold the machine before it is replaced. State clearly
assumptions made. [JNTU (Hech.) 93, Madras B.E. (Mech) 1984]
Answer: Nineteen years

7. A truck has been purchased at a cost of Rs. 1,60,000. The value of the truck is
depreciated in the first 3 years by Rs. 20000 each year and Rs. 16000 per year
thereafter. Its maintenance and operating cost for the first three years are Rs.
16000,18000 and 20000 in that order and then increase by Rs. 4000 every year.
Assuming an interest rate of 10%. Find the economic life of the truck.
[JNTU (Mech.) 2001/CCC, JNTU (ECE) CCC 2001/43301]

Answer: 7 years
540 Operations Research : Theory and Practice

8. A truck owner estimates that the running costs and the salvage value of trucks
for various years will be as illustrated below. If the purchase price of a truck is
Rs. 80000 estimate the optimum replacement age for the truck. Take money's
value as 15% per annum.

Year 1 2 3 4 5 6 1 8
Running Cost (Rs) 6000 7000 9000 12000 15000 20000 25000 30000
Resale Price (Rs) 60000 40000 30000 30000 25000 20000 20000 20000
[JNTU (CSE) 98]

Answer: By the end of 5 th year

9. If you wish to have a return of 10% per annum on your investment. Which of
the following plans would you prefer.
Plan ,4 Plan£
First cost 2,00,000 2,50,000
Scrap value after 15 years 1,50,000 1,80,000
Excess of annual revenue over 25,000 30,000
manual disbursement

Answer: Plan A

10. Singareni Collieries Company Ltd (SCCL) has received offers for two types of load
hauling dumpers A and B. A has a pay load of 25 tonnes and is priced at Rs.
4,00,000 while B also with pay load of 25 tonnes, is priced at Rs. 3,60,000. The
operating costs over the estimated life of 5 years for both the types are as follows.

Year 1 2 3 4 5
Type^(Rs) 8000 9000 10000 11000 12000
Type5(Rs) 14000 16000 18000 20000 22000

Which type of dumper is to be preferred.

Answer: Type B

11. A manual stamper currently valued at Rs. 1000 is expected to last two years and
costs Rs. 4000 per year to operate. An automatic stamper which can be
purchased for Rs. 3000 will last for 4 years and can be operated at an annual
cost of Rs. 3000. If money carries a rate of interest of 10% per annum, determine
which stamper should be purchased. [Bangalore B.E. (Mech) 1982]

Answer: Purchase an automatic stamper

Replacement Analysis 541

12. There are two machines^ and B. A is replaced after every 3 year where as B is
replaced after every 8 years. The yearly cost of the machines are given below :
Year 1 2 3 4 5 6
A 1000 200 400 1000 200 400
B 1700 100 200 300 400 500

Assuming the value of money to be 10% per year, determine which machine
should be purchased. [JNTU (CSE) 95/S]

Answer: Considering six years period for both the machines, machine A appears
to be cheaper than B. Hence machine A should be purchased.
13. The cost of a new car is Rs. 10,000. Compare the optimum moment of
replacement assuming the following cost information :

Assuming that repairs are made at the end of each year only if the car is to be
retained and are not necessary if the car is to be sold for its salvage value. Also
assume that the rate of discount is 10%.
Age of car Repair costs / year in Rs. Salvage value at the end of the year in Rs.
1 5000 8000
2 10000 6400
3 10000 5120
[JNTU (CSE) 93/S]

Answer: The car must be replaced at the end of first year.

14. Let v = 0.9 and initial price is Rs. 5000. Running cost varies as follows:
Year 1 2 3 4 5 6 7
Running cost in (Rs) 400 500 700 1000 1300 1700 2100

What would be the optimum replacement interval?

Answer: Six years

15. A pipeline is due for repairs. It will cost Rs. 10000 and last for 3 years.
Alternatively a new pipeline can be laid at a cost of Rs. 30000 and last for 10
years. Assuming cost of capital to be 10% and ignoring salvage value, which
alternative should be chosen? [CJL (May) 1981]

Answer: Existing pipe line should be continued

542 Operations Research : Theory and Practice

16. The cost pattern of two machines Mx andAf2 when money value is not
considered is
Cost at the beginning of the year in (Rs)
Year Mx M2

1 900 1400
2 600 100
3 700 700

Find the cost pattern for each machine when money is worth 10% per year and
hence find the machine which is less costly.

Answer; Machine B should be purchased

17. An individual is planning to purchase a car will cost Rs. 1,20,000. The resale
value of the car at the end of the year is 85% of the previous year value.
Maintenance and operation costs during the first year are Rs. 20,000 and they
increase by 15% every year. The minimum resale value of car can be Rs. 40,000.

(i) When should the car be replaced to minimise average annual cost (ignore
initial)?

(ii) If interest of 12% is assumed, when should the car be replaced?

Answer: At the end of 1st year
18. The cost of a new machine is Rs. 5000 the maintenance cost of nth year is given
by Rn = 500 (n - 1), n = 1, 2 . . . suppose that the discount rate per year is 0.5.
After how many years it will be economical to replace the machine by a new
one? [JNTU (CSE) 2001]

Answer; 5 years

19. The annual costs of two machines Xand Y are given below. Find their cost
patterns if money value is 10% per year, what conclusion do you derive about
which machine is economical?

Year 1 2 3
Machine X(Rs) 1900 1100 1400
Machine F(Rs) 2700 300 1400

Answer: Machine X is more economical.

Replacement Analysis 543

20. The initial cost of an item is 15000 and maintenance and running costs in
rupees for different years are given below :
Year 1 2 3 4 5 6 7
Running cost 2500 3000 4000 5000 6500 8000 10000

What is the replacement policy to be adopted if the capital is worth 10% and
there is no salvage value.
Answer: Optimal replacement period = 5 years ( Rs. 7609).

21. A special lathe machine costs Rs. 1 lakh, annual operation and machine cost is
Rs. 1,000 and increases at a rate of Rs. 750 for 2nd and 3rd year and at a rate
of Rs. 1250 for the remaining life. The life of equipment is 10 years. Salvage
value at the end of the first year is Rs. 80,000 and falls at a rate of 10,000 every
year for the first 4 years. Afterwards it has no solvage value. What should be
economic life of lathe if the rate of return on capital invested is 25%.
Answer: At the end of 2nd year (Rs. 18000) [JNTU - CSE 97]

10*6 Group Replacement

Group Replacement of Items That Fail Completely :
This policy is concerned with the items that either work perfecdy, or work
partially, or inefficiently or fail completely. This situation generally happens when the
system consists of a large number of identical low cost items that are increasingly liable
to failure with age. In such cases, the replacement of individual items would incur a
set of costs, which is independent of the number replaced. However it may be
advantageous to replace Jl the items at a time at a fixed interval. This policy is known
as group replacement policy and is very attractive, particularly when
1. The value of any individual item is so small.
2. The cost of keeping records of individual ages is high that cannot be justified.
3. The purchase of such identical items in bulk can be had at discounted rate.
4. Average individual replacement would be costlier than the average group
replacement.
5. If sufficient number of standby machines are available.
6. New designs of the equipment considerably increase the production rate.
In all the above cases the two types of replacement policies considered are :
1. Individual Replacement: Under this policy, an item is replaced immediately
after its failure.
2. Group Replacement: Under this policy a decision will be taken so as to replace
all the items irrespective of the fact that the items have failed - not failed,
provided if any item fails, before the optimal time it may be replaced
individually.
544 Operations Research : Theory and Practice

Algorithm for Deciding Group Replacement Policy :

Step 1: Find the probability of failure of items at the end of each period.
Step 2 : Find the number of replacements made at the end of each period with reference
to probability of failures at the end of each period considering its previous
replacements.
Step 3 : Calculate cost of individual replacement at the end of each period.

Step 4 : Calculate cost of group replacement at the end of each period.

Step 5 : Calculate total cost of group replacement including individual replacements by

adding step 3 and step 4.

Step 6: Calculate the average cost per period by dividing the result in step 5 with period
number.

Step 7: Identify the least among the average cost per period as the period of group
replacement policy.
ILLUSTRATION - 7

1000 bulbs are in use and it costs Rs. 10 to replace an individual bulb which has
burnt out If all bulbs were replaced simultaneously it would cost Rs, 4 per bulb.
It is proposed to replace all bulb fs atfixed intervals of time, whether or not they
have burnt out and to continue replacing burnt out bulbs as and when they fail
The failure rates have been observed for certain type of light bulbs are as
follows:

Week 1 2 3 4 5
Percent failing by the end of week 10 25 50 80 100

At what intervals all the bulbs should be replaced? At what group replacement
price per bulb would a policy of strictly individual replacement become
preferable to the adoptedpolicy. [ JNTU CSE 2001]
! Solution ;

Step 1: To find out the probability of failure of items at the end of each week.

The probability of failure of light bulbs in first week

The probability of failure of light bulbs in second week.

D (25-10)
= 0.15
100
Replacement Analysis 545
The probability of failure of light bulbs in third week

The probability of failure of light bulbs in fourth week

The probability of failure of light bulbs in fifth week

(100-80)
-p&- 100 - 0 2 0

Sum of all probabilities is 1.

i.e., Pl+P2+^ + P4 + P5=l
All further probabilities P 6 , P 7 , P g and so on will be zero.

Step 2: Calculation of number of replacements made considering previous

replacements,

Let N{ be the number of replacements made at the end of itn week, if all 1000
bulbs are new initially.
Thus

Ni = NOPX = 1000 x 0.1 = 100

N2=N0P2 + Nl PY = 1000 x 0.15 + 100 x 0.10 = 160
l = 1000x0.25+ 100x0.15+ 160x0.10 = 281
4 0 4 l s 2 2 +N$Pl = 377
N5=N0P5+Nx P4 + N2PS + NSP2+N4PX = 350
N6 = 0 + NxP5 + N2P4 + NsP$ + N4P2 + N5PX = 230
N>j = 0 +

From above results it is clear that number of bulbs burnt out increases upto
fourth week and decrease upto sixth week and again start increasing. The whole
system comes to a steady state where the proportion of bulbs failing in each week is
the reciprocal of their average life.

As the mean age of bulbs.

= 1 x />! + 2 x P2 + 3 x Pg + 4 X P4 + 5 x P 5
= 1 x 0.1 + 2 x 0.15 + 3 x 0.25 + 4 x 0.30 + 5 x 0.20= 3.35 week.
546 Operations Research : Theory and Practice

.'. number of failures in each week in steady state become = . or = 299

O.OD

.-. cost of replacing bulbs individually only on failure = 10 x 299 = Rs. 2990.
Step 3 : Calculating the cost of individual replacement at the end of each period cost
of individual replacement at
end of first week = 100 x 10 = 1000
end of second week = 1 6 0 x 1 0 = 1 6 0 0
end of third week = 281 x 10 = 2810
end of fourth week = 377 x 40 = 3770

Step 4 : Calculating the cost of group replacement at the end of each period.
end of first week = 1000 x 4 = 4000
end of second week = 4000 + 1000 = 5000
end of third week = 5000 + 1600 = 6600
end of fourth week = 6600 + 2810 = 4410
Step 5 : Calculating total cost of group replacement including individual replacement
i.e., adding values of step 3 and step 4.
end of first week = 4000 + 1000 = 5000
end of second week = 5000 + 1600 = 6600
end of third week = 6600 + 2810 = 9410
end of fourth week = 9410 + 3770 =13180

Step 6 : Calculate average cost per week

end of first week = ^ y ^ = 5000

y
end of second week = —-— = 3300

end of third week = ^—» = 3136.67

3
1 Q I OA
end of fourth week = — - — = 3295
4
Step 7 : To identify the least among average cost per period.

It is identified as third week, i.e., 3136.67, so it would be optimal to replace all

the bulbs after every 3 weeks, other wise the average cost will be increasing.
Replacement Analysis 547

LLUST-RATION - 8 —
A factory has a large number of bulbs all of which must be in working condition. The
mortality of bulbs is given in the following table :
Week Proportion of Bulbs Failing
During the Week
1 0.1
2 0.15
3 0.25
4 0.35
5 0.12
0.03

If a bulb fails in service, it costs 3.50 to replace but if all bulbs are replaced at a time
it costs Rs. 1.20 each. Find the optimum group replacement policy. (Assume 1000 bulbs
as available in the beginning). [JNTU (Mech.) 99]

Solution :
No = 1000
ATj =NOPX= 1000x0.1 = 100 bulbs
N2=NlPl+N0P2
= 100x0.10+ 1000x0.15
= 160 bulbs

= 160 x 0.01 + 100 x 0.15 + 1000 x 0.25

= 281 bulbs
AT4= NsPx + N2P2 + NX Ps + N0P4
= 281 x 0.1 + 160 x 0.15 + 100 x 0.25 + 1000 x 0.35
= 427
N5 =7V 4 P 1+ Ar 3 /> 2 + iV2Jp3 + iV 4 P 4 + Ar0P5
= 427x0.1 + 281x0.15 + 160x0.25 + 100x0.35 + 1000x0.12
= 279.

= 279x0.1 +427x0.15 + 281x0.25 +160 x 0.35 + 100x0.12 +

1000x0.03.
= 260 bulbs.
548 Operations Research : Theory and Practice

End of the Total no. of Cummulative Cost of Cost of group Total cost Average cost
week bulbs failed no. of failure replacement replacement per week
(Rs. 3.50) (Rs. 1.20)
1 100 100 350 1200 1550 1550
2 160 260 910 1200 2110 1055
3 281 541 1893.5 1200 3093.5 1031.1
4 427 968 3388.5 1200 4588 1147
5 279 1247 43G4.5 1200 5564.5 1113
6 260 1507 5274.5 1200 6474.5 1079.08

/. bulbs should be replaced by the end of every 3 week.

Practice Problems
1. Find the cost per period of individual replacement policy of an installation of
300 lighting bulbs, given following :
(i) Cost of replacing individual bulbs is Rs. 3.
(ii) Conditional probability of failure is given below :

Week No. 0 1 2 3 4
Conditional probability of failure 0 J_ 1 2 1
10 3 3

JNTU (CSE) 95]

Answer: Expected life of each light bulb is 3 week and the average cost of individual
replacement of 300 light bulbs is Rs, 206.

2. An electric company which generates and distributes electricity conducted a study

on the life of poles. The appropriate life data are given in the following table :

Year after installation 1 2 3 4 5 6 7 8 9 10

Percentage poles failing 1 2 3 5 7 12 20 30 16 4

(i) If the company now installs 5000 poles and follows a policy of replacing
poles only when they fail, how many poles are expected to be replaced
each day during the next ten year.

(ii) If the cost of replacing individually is Rs. 160 per pole and if we have a
common group replacement policy, it costs Rs. 30 per pole. Find out the
optimal period to group replacement. [JNTU (ECE) 95/CCC]

Answer : (i) 5533 poles (ii) All poles after 6th year
Replacement Analysis 549

3. The following failure rates have been observed for a certain type of light bulbs.
End of week 1 2 3 4 5 6 7 8
Probability of failure to date 0.05 0.13 0.25 0.43 0.68 0.88 0.96 1.00

The cost of replacing an individual bulb is as 2.25, the decision is made to

replace all bulbs simultaneously at fixed intervals and also to replace bulbs as
they fail in service. If the cost of group replacement is 60 paise per bulb and
the total number of bulbs is 1000. What is best interval between group
replacement? [0U (Mech.) 88]

Answer: Replacement at the end of third week.

4. Suppose that a special purpose type of light bulb never lasts longer than two
weeks. There is chance of 0.3 that a bulb will fail at the end of first week. There
are 100 new bulbs initially. The cost per bulbs for individual replacement is Rs.
1.25 and the cost per bulb for a group replacement is Re. 0.50.
Is it cheapest to replace all bulbs;
(i) Individually. (ii) Every week.
(iii) Every second week. (iv) Every third week. [OU - Mech 89]
Answer: Individual replacement.
6. The probability Pn of failure just before age n are shown below. If individual
replacement costs Rs. 1.25 and group replacement costs Re. 0.50 per item, find
the optimal group replacement policy.

n= l 2 3 4 5 6 7 8 9 10 11
0.01 0.03 0.05 0.07 0.10 0.15 0.2 0,15 0.11 0.08 0.05

[JNTU (Mech.) 87]]

Answer: Replacement after every 5 weeks.
6, It has been suggested by a data processing firm that they adopt a policy of
periodically replacing all the 1000 tubes in a certain piece of equipment. A
given* type of tube is known to have a mortality distribution (probability of
failure) shown in the following table.
Tube failure / week 1 2 3 4 5
Probability of failure 0.3 0.1 0.1 0.2 0.3

The cost of replacing the tubes on an individual basis is estimated to be Re.

1.00 per tube. The cost of group replacement policy average Rs, 0.30 per tube.
Compare the cost of preventive replacement with that of remedial replacement.
[JNTU (Mech.) 91, 0U (EEE) 87]
Answer: It is optimal to have a group replacement after every 4 week along with
the individual replacement.
550 Operations Research : Theory and Practice

7. A decorative series lamp set circuit contains 10,000 bulbs, when any bulb fails
it is replaced and the cost of replacing a bulb individually is Re. 1 only. If all
the bulbs are replaced at the same time the cost per bulb would be reduced to
35 paise. T h e percent surviving say s(t) at the end of month t, are given as

97 90 70 30 15 0

Determine the optimal replacement policy. [ONTU (ECE) 93/S CCC]

Answer : After 3rd month
8. A computer has a large number of electronic tube, that are subject to mortality
as given below :
Period Age of failure (hours) Probability of failure
1 0-100 0.10
% 101-2 0.26
3 201-300 0.35
4 301-400 0.22
5 401-300 0.07
If the tubes are group replaced, the cost of replacement is Rs. 15 per tube.
Group replacement can be done at fixed intervals in the night shift when the
computer is not normally used. Replacement of individuals tubes which fails in
services costs Rs. 60 per tube. How frequently should the tubes be replaced?
[JNTU (CSE) 92/S]

Answer: Minimum cost for group replacement is Rs. 18600 for an interval of 200
hours.
9. There are 1000 bulbs in the system survival rate is given below :
Week 0 1 2 3 4
Bulb in operation at the end of week 1000 850 500 200 100

The group replacement of 1000 bulbs costs Rs. 100 and individual replacement
is Re. 0.50 per bulb. Suggest suitable replacement policy. [JNTU (Mech.) 90]

Answer: One week group replacement policy ; Average cost = Rs. 175 per week;
Cost of individual replacement / week = Rs. 196.
10. The following mortality rates have been observed for a certain type of light bulbs.
Week 1 2 3 4 5
Percent failure by end of week 10 35 58 80 100
There are 1000 bulbs in use and it cost Rs. 2 to replace an individual bulb which
has burnt out. If all bulbs-were replaced simultaneously it would cost 0.50 paise
per bulb. It is proposed to replace all bulbs at fixed intervals whether or not
they have burnt out and to continue replacing burnt out bulbs as they fail. At
what interval should all bulbs be replaced. [Bangalore B.E. (Mech) 1984
JNTU (FDH, Hech.) 92]
Replacement Analysis 551

11. In a machine shop, a particular cutting tool cost Rs. 6 to replace. If a tool breaks
on a job, the production disruption and associates with amount to Rs. 30. The
past life of a tool is given as follows :
Job Number 1 2 3 4 5 6 7
Proportion of broken tools 0.01 0.03 0.04 0.13 0.25 0.55 0.95
[I.A.S 1989]
After how many jobs should the shop replace a tool before it breaks down.

Answer : Group replacement after a week.

12. The following mortality rates have been found for a certain type of coal cutter
motor.
Weeks 10 20 30 40 50
Total %failure upto end of 10 weeks 5 15 35 65 100

If the motors are replaced over the week, the total cost is Rs. 200. If they fail
during the week the total cost is Rs. 100 per failure. Is it better to replace the
motors before failure and if so when?
Answer: Motors should be replaced every 20 weeks.
13. The mortality rates as obtained for an electronic component are noted in the
table below :
Month 1 2 3 4 5 6
Percentage fails at the end of the month 8 22 45 70 85 100

There are 1500 items in operation. It costs Rs. 20 to replace an individual item
and Re. 0.50 per item if all items are replaced simultaneously. It is decided to
replace all items at fixed intervals and to continue replacing individual item as
and when they fail. At what intervals should all items be replaced?
[JNTU (Mech.) 91]

14. The following failure rates have been observed for certain types of light blubs

Week 1 9 3 4 5
Percent failure by end of week 10 25 58 80 100

There are 1000 bulbs in use, and it cost Rs. 2 to replace an individual bulb which
is burnt out. If all bulbs were replaced simultaneously it would cost 50 paise per
bulb. It is proposed to replace all bulbs at fixed intervals, whether they have
burnt out or not, and to continue replacing bulbs on and when they fail. At
what intervals, all the bulbs should be replaced? [JNTU - CSE/ECE 2001]
552 Operations Research : Theory and Practice

Review Questions
1. What are the costs involved in failure and replacement analysis of equipment.
Explain.
2. Discuss the effect of running age of machine on failures and machine behaviour
with the aid of a graph (Bath Tub Curve). Where do you fit machine
replacement in the graph.
3. Explain Machine life Cycle, with reference to costs.
4. How does an equipment behave in the following stages? What type of
maintenance do suggest?
(a) Early stage when just bought and installed.
(b) When it is producing at its highest rate.
(c) When its expected life is almost completed.
5. When do you recommend replacement of a machine with a new one?
6. Discuss the effect of age on machine resale value and running cost with an aid
of a graph.
7. Discuss the replacement policy when money value is not changing with time.
8. Write a note on replacement policy when money value is changing with time.
9. What do you understand by group replacement? When do you opt this policy?
10. Discuss the step by step procedure to decide the interval of group replacement.
11. Write a short note on group replacement with examples of its application.
12. Distinguish between the individual replacement and group replacement
policies.
13. What policy do you recommend for the items that fail completely. Justify your
answer with examples.
14. Discuss the use of probabilities in replacement policies.
15. Distinguish between repairable and non-repairable system with reference to
replacement policies.
16. Discuss the importance of group replacement in the case of non-repairable
systems.
17. Discuss different types of failures that occur on a machine.
18. Distinguish between gradual failures and sudden failures and their effects.
19. Discuss the following failures and suggest the maintenance policies to be
adopted.
(a) Early failures or infant failures.
(b) Random failures or chance failures.
(c) Wear-out failures or old-age failures.
20. Discuss the effect of money value on machine replacement.
Replacement Analysis 553

Objective Type Objectives

1. When a machine is in infant stage, the major maintenance policy considered is
(a) corrective maintenance (b) preventive maintenance
(c) contractual maintenance with OEM (d) condition monitoring

2. When money value changes with time at 10% then PWF for first year is
(a) 0.909 (b) 1 (c) 0.826 (d) 0.9

3. Which of the following maintenance policy is not used in old age stage of a
machine
(a) operate to failure and corrective maintenance
(b) renovation/reconditioning
(c) replacement
(d) scheduled preventive maintenance
4. When money value is changing with time @ 20%, the discount factor for 2nd
year =
(a) 1 (b) 0.833 (c) zero (d) 0.6944
5. If the probability of failure of a machine is gradually decreasing, the failure
mode is said to be
(a) progressive (b) retrogressive
(c) regressive (d) recursive
6. Which of the following replacement policy is probability model
(a) when money value does not change with time where time is continuous
variable
(b) when money value does not change with time where time is a discrete
variable
(c) when money value changes with time
(d) group replacement policy
7. Which of the following replacement policy is considered to be dynamic
(a) when money value does not change with time where time is continuous
variable
(b) when money value does not change with time where time is descrete variable
(c) when money value changes with time
(d) none of the above
8. A machine is to be replaced if the average running cost
(a) is not equal to current running cost
(b) of current period is greater than that of next period
(c) till current period is greater than that of next period
(d) of current period is less than that of next period.
554 Operations Research : Theory and Practice

9. Which of the following is correctl; assumption for replacement policy when

money value does not change with time

(a) no initial cost (b) scrap value is zero

(c) scrap value is constant (d) no maintenance cost

10. In replacement analysis the maintenance cost is the function of

(a) present value (b) time

(c) maintenance policy (d) resale value

11. Find the oddman out

(a) present worth factor (PWF) (b) discounted rate(DR)
(c) depreciation value (DV) (d) mortality tables

12. Which of the following is not used in group replacement policy decisions

(a) failure probability (b) loss due to failure

(c) cost of individual replacement (d) present worth factor

13. Reliability of an item is

(a) failure probability (b) 1 - failure probability

(c) 1/failure probability (d) warranty period

14. Group replacement policy is applicable for

(a) repairable items (b) items that fail partially

(c) items that fail completely (d) dissimilar items

15. The maintenance cost are assumed to be mostly dependent on

(a) calendar age (b) running age

(c) user's age (d) manufacturer's age

16. Group replacement is not preferred for

(a) large number of identical items

(b) low cost items where record keeping may be costly or difficult
(c) items that fail completely
(d) repairable items
Replacement Analysis 555
17. Which cost of the following is irrelevant to replacement analysis

(a) capital (purchase) cost of machine

(b) operating cost of machine
(c) production cost of the machine
(d) maintenance cost
18. Group replacement policy is most suitable for
(a) blood group (b) street lights
(c) truck (d) any of the above

19. Replacement problem arises when

(a) old item becomes too expensive to operate or maintain

(b) old item completely fails to work
(c) a better design or efficient model is available that can increase rate of
production
(d) all the above

20. The type of failure that usually occurs in old age machine is

(a) early failures (b) random failure

(c) wear-out failures (d) chance failures

21. Decreasing failure rate is usually observed in stage of the machine

(a) infant (b) youth

(c) old age (d) any

22. Chance failures that occur on a machine are commonly found on a graph of
time (X-axis) Vs failure rate (F axis) as
(a) hyperbolic
(b) parabolic
(c) straight line nearly parallel to X axis
(d) straight line nearly parallel to Y axis

23. The curve used to interpret machine life cycle is

(a) learning curve (b) bath tub curve

(c) time series (d) ogive curve
556 Operations Research : Theory and Practice

24. If machine becomes old* the failure rate expected will.be

(a) decreasing (b) constant
(c) increasing (d) may increase or decrease
25. The production rate on a machine is expected to be highest or peak in
stage
(a) infant (b) youth (c) old age (d) any

26. A replacement decision is common in stage

(a) youth (b) infant (c) old age (d) none

27. The preventive maintenance is expected to enhance of the

machine
(a) infancy (b) youth (c) old age (d) no stage
28. The replacement policy that is imposed on an item irrespective of its failure
(i.e. it is replaced even though it has not failed) is
(a) individual replacement (b) group replacement
(c) if money value is constant (d) if money value varies with time
29. The replacement model that uses failure probability is
(a) individual replacement
(b) group replacement
(c) replacement when money value is constant
(d) replacement when money value is time dependent
30. Replacement stage is said to be reached if is increasing
(a) failure rate (b) failure cost
(c) failure probability (d) any of these

Fill in the Blanks

1. When money value changes with time @ 20%, then present worth factor (PWF)
during first year is taken as

2- When money value changes with time @ r then discount factor for nth year -

3. There are two machines whose costs are calculated as follows.

Machine - 1 : Discounted @ 10% twice a year
Machine - 2 : Discounted @ 20% per annum,
which machine is preferable
Replacement Analysis 557

4. In retrogressive failures, the failure probability with time

5. The probabilistic model among replacement policies is

6. The replacement policy when money value does not change with time can be
placed under models of OR classification

7. A machine is to be replaced if the average running cost of current period is

to that of period

8. Scrap value is assumed to be for replacement policy when

money value does not change with time

9. The maintenance cost is function of

10. If failure probability of machine is / , then its reliability is _

11. The replacement policy applicable for the items that completely fail is _

12. The suitable replacement policy applicable for low cost, identical, large number
of items is

13. The wear-out failures usually occur in stage o machine

14. The highest and most efficient production rate is found in stage
of the machine life cycle

15. The replacement decision is a major concern for machine in stage

of machine life cycle

16. The shape of graphical representation of machine life cycle is

17. The shape of early failure rate in a machine life cycle graph is
18. The preventive maintenance is expected to increase the life of machine in
stage of the machine

19. The discount factor at the end of* 1st year when money value changes @ 10%
is

20. Contractual maintenance with original equipment manufacturer (OEM) is

preferred in stage of the machine
558 Operations Research : Theory and Practice

Answers
Objective Type Questions;

Me) 2. (b) 3. (d, 4.(b) 5.(b) .

6.(d) 7. (c) 8. (d) 9.(c) 10. (b)
11. (d) 12. (d) 13. (b) 14. (c) 15. (b)
16. (d) 17. (c) 18. (b) 19. (d) 20. (c)
21- (a) 22. (c) 23. (b) 24. (c) 25. (b)
26. (c) 27. (b) 28. (b) 29; (b) 30. (d)
Fill in the Blanks :
1. 1 2.(l+r)-"+1
3. machine -1 4. decreases
5. group replacement 6. static
7. less than or equal to (2), next 8. constant
9. running time or running age of machine 10. 1-/,
11. group replacement 12. group replacement
13. old age 14. youth
15. old age 16. bath tub
17. decreasing hyperbolically 18. youth
19. 0.909 20. Infant or early
560 Operations Research : Theory and Practice

CHAPTER AT A GLANCE
Start

Write the payoff matrix

Apply maximm-minimax principle

Identify the value of game and write the Yes

optimal strategy of the players

Solve by using analytical (or algebric) or

matrix method for mixed strategic games

No
Use dominance rule to reduce the size
of the payoff matrix to either 2*2
2xw ormx2, size (order)

Yes

Yes
Use graphical method to solve the problem.

Formulate and solve as a LP problem

Stop
Game Theory 561

1L0 Introduction
In the competitive world, it is essential for an executive to study or at least guess
the activities or actions of his competitor. Moreover, he has to plan his course of actions
or reactions or counter actions when his competitor uses certain technique. Such war
or game is a regular feature in the market and the competitors have to make their
decisions in choosing their alternatives among the predicted outcomes so as to
maximise the profits or minimising the loss.

Hie game theory has taken its importance in the business management in 1944,
when Von Neumann published "Theory and practice of Games and Economic
Behaviour".

11.1 Theory of Games: The Terminology Used

A competitive situation is called game, in this context.

11.1.1 Properties of a Game

1. There are finite number of competitors called 'players'.

2. Each player has a finite number of possible courses of action called 'strategies'.
3. All the strategies and their effects are known to the players but player does not
know which strategy is to be choosen.

4. A game is played when each player chooses one of his strategies. The strategies
are assumed to be made simultaneously with an outcome such that no player
knows his opponents strategy until he decides his own strategy.
5. The game is a combination of the strategies and in certain units (usually in
terms of money) which determines the gain (shown in positive figures) or loss
(shown in negative figures).

6. The figures (either gain or loss) shown as the outcomes of strategies in a matrix
form is called "pay-off matrix"

7. The player playing the game always tries to choose the best course of action
which results in optimal pay off, called "optimal strategy11.
8. The expected pay off when all the players of the game follow their optimal
strategies is known as "Value of the game". The main objective of a problem of
games is to fmd the value of the game.

9. The game is said to be 'fair game if the value of game is zero, otherwise it is
known as 'unfair.
562 Operations Research : Theory and Practice

11*1.2 Definitions of Terms Used

1. Startegy: A startegy for a player has been defined as a set of rules or alternative
courses of action available to him in advance, by which player decides the course
of action that he should adopt.Startegy may be of two types :
(a) Pure Strategy : If the players select the same strategy each time, then it is
referred to as pure-strategy. In this case each player known exactly what
the other is going to do i.e., there is a deterministic situation and the
objective of the players is to maximize gains or to minimize losses.
(b) Mixed Strategy : When the players use a combination of strategies and
each player always kept guessing as to which course of action is to be
selected by the other player at a particular occasion then this is known as
mixed-strategy. Thus, there is a probabilistic situation and objective of the
player is to maximize expected gains or to minimize losses. Thus, mixed
strategy is a selection among pure strategies with fixed probabilities.
2. Optimum Strategy : A course of action or play which puts the player in the
most preferred position, irrespective of the strategy of his competitors is called
an optimum strategy. Any deviation from this strategy results in a decreased
pay-off for the player.
3. Value of the Game : It is the expected pay-off of play when all the players of
the game follow their optimum strategies. The game is called/air if the value
of the game is zero and unfair if it is non-zero.
4. Two-Person Zero-Sum Game : There are two types of two person zero-sum
games. In one, the most preferred position is achieved by adopting a single
strategy and therefore the game is known as the pure strategy game. The second
type requires the adoption by both players a combination of different strategies
in order to achieve the most preferred position and is, therefore, referred to as
the mixed strategy game.
5. Payoff Matrix : A two-person zero-sum game is conveniently represented by a
matrix as shown below. The matrix, which shows the outcome of the game as
the players select their particular strategies, is known as the payoff matrix. It is
important to assume that each player knows not only his own list of possible
courses of action but also of his opponent.
Let player A have m courses of action {A^.A^, . . . A^ and player B has n courses
of action (Bx, 3%, . . . Bn). The number n and m need not be equal. The total number
of possible outcome is, therefore (m x n). These outcomes are shown in the following
table.
A }s payoff matrix B 's payoff matrix
Player B Player B
B
\ B2 B
n
au a
\2 ~a\\ ~an ~a\n
Player^ a21 a22 Player^ -an

a a
m\ nm a
m2
Game Theory 563

11.2 Types of Games

Games can be classified into different categories with reference to various
aspects as given below.

11.2,1 With Reference to Number ofPlayers

With reference to the number of players playing the game, the games are
divided into two categories Viz. (1) Two-person games and (2) Multi-person games.
1. Two Person Games : If only two players are playing the game then the game
is said to be two person game. The rows and column in the pay off matrix
represent the strategies of these two players.
e.g.: (a) Chess players
(b) Marketing strategies of Coke and Pepsi in India.
2. Multi-person Game : It is a game in which more than two persons play.
e.g. : Competition among tooth paste producers, soap producers.

11.2,2 With Reference to Nature of Out Comes

1. Zero-sum game. 2. Non-zero sum game.
1. Zero-sum Game : If loss of a player is gain of other player and vice-versa, then
the game is called zero-sum game.
2. Non-zero Sum Game : In contrast to the above, if a gain of player is not a loss
of another player (positive sum game) or loss of a player is not a gain of another
player (negative sum game), then such games are non-zero sum games.

e.g.; Suppose there are two players, say Colgate and Pepsodent. In a colony
there are 1000 people who either use Colgate or Pepsodent. Now, if say
20 people have changed their brand from Colgate to Pepsodent due to
effective advertisement then the loss of Colgate is gain for Pepsodent, This
is a zero-sum game.
If a group of 30 people has stopped using any toothpaste and use
neem-sticks. Here loss of any player is not a gain to other player. This
game is non-zero sum (negative sum) game.

1L2.3 With Reference to Certainty

1. Deterministic games. 2. Probabilistic games.
1. Deterministic Games : If the game yields a solution with one certain, single
strategy (pure strategy) for each player, the game is deterministic. In this game
the strategies are pure and saddle points exist.

2. Probabilistic Games : If a player adopts more than one strategy with some
probabilities or prepares a blend of two or more strategies, these games are
said to be probabilistic. Here, mixed strategies are used as no saddle points
exist.
564 Operations Research : Theory and Practice
112.4 With Reference to Value of the Game
1. Fair game 2. Unfair game

1. Fair Game : If the value of the game is zero, i.e., neither player wins nor loses
(drawn), the game is said to be fair.

2. Unfair Game : If one of the player wins (other loses), or the value of the game
is non zero (may be positive or negative), it is unfair.

The following is the classification of games shown with a chart.

Fair
s Deterministic <T
Zero-sum./'^ Unfair
games ^C.
Fair
Two-person Probabilisitic < ^
games
Unfair
Non-zero-sum
games

Multiperson
games Same as above

FIGURE I I.I : CLASSIFICATION OF GAMES

11.3 Two Person-zero Sum Games or Rectangular Games

A game with only two persons with a condition that the loss of one player is gain
of the other and vice-versa (so that total sum is zero) is said to be two-person zero-sum
game or rectangular game.

Pay Off Matrix : The qualitative measures of returns of the players in terms of
gains or losses when players select their particular strategies can-be represented in
the form of a matrix in m rows and n .columns, called pay off matrix. The m rows are
the m courses of action of one player while n columns represent the strategies of the
other player. Thus one player say A has m strategies^, A%9 A% . . .Am and player B
has n strategies as Bp B^ B§ . . . Bn. When each player chooses one of his strategies
say Ai and £ -, then the expected out come x^ is represented as the elements of the
matrix. It is a general convention to read the pay off matrix with reference to row
player (here A) always.Thus a positive sign of x- is gain to A or loss to B while the
negative sign of x^ (i.e., - x-) is loss of A or gain of B.
Game Theory 565

Obviously player */!' wishes to maximise his gains while player B wishes to
minimise his loss.

An exemplary pay of matrix of two-person zero-sum game is shown below.

(xj i is pay off when A uses A j strategy and B uses B^ strategy x 12 is pay off when
A applies A y and B applies B% strategies and so on).
Player Bys strategies

*11 x 12 x13 X
\n
x X X
\2 32 2n
x
Player ^4*5 strategies n

11.3-1 Assumptions of the Game

1. Each player has finite number of possible courses of actions (strategies) with
him.

2. The list of strategies of each player need not be same.

3. Player in Rows attempts to maximise his gains while player in columns tries to
minimise his losses.
4. The decision of both players are made individually prior to the play with out
any communication between them.
5. The decision is supposed to be made simultaneously and also announced
simultaneously so that neither player has any advantage over the other due to
the direct knowledge of the other player's decision.
6. Both players know their own pay ofrs as well as the pay off of each other.

7. The positive sign of the pay off indicates the gain to row player or loss to column
player and negative sign indicates loss to row player or gam to column player.

11.3.2 Solving Two Person-zero Sum Games

Two person zero sum games may be deterministic or probabilistic. The
deterministic games will have saddle points and pure strategies exist in such games.
In contrast, the probabilistic games will have no saddle points and mixed strategies
are taken with the help of probabilities. These are explained in following sections.
566 Operations Research : Theory and Practice

Games With Saddle Points - Deterministic Games - with Pure Strategies :

These games can be solved by using one of the following methods.
1. Minimax and Maximin principle.
2. Dominance principle.
Minimax and Maximin Criteria :
In a given pay off matrix of two players, it is evident that the row player will
maximise his minimum profits, thus applies maxi-min principle. Similarly, the
column player attempts to minimise his maximum loss using his mjM-max rule.
For example for player^ (in rows), minimum value in each row represents his
least gain when he chooses to play that particular strategy. In other words, it is the
minimum gain he can be confident, what so ever the strategy that his opponent uses.
Thus in each row, when such minimal values are calculated, among all these least
gains, he wishes to select the maximum of these values. This selection of maximum
gains (largest value) among the row minimum values is referred to as Maxi-Min
principle.
Similarly, for player in columns (say B) who is assumed to be loser, the
maximum value of each column denotes the maximum loss to him if he adopts that
particular columns strategy. In other words, it is the worst case of B for which he must
be prepared. These are written beneath the pay off matrix called column maxima
from which he can choose least (to minimise his losses). This is called Mini-max
principle. If maximum value of row minima i.e., maximin value of row player is equal
to the minimax value of column player (i.e., minimum value of column maxima), then
the game is said to have 'Saddle (Equilibrium) Points', and the corresponding strategies
are said to be 'Optimal Strategies'. The value of the pay off (say V) at this saddle
(equilibrium) point is known as 'Value of the Game'. A game may have more than one
saddle point but the value will be same.
Steps To Find Saddle Points by Applying Minimax-Maximin Criteria :
Step 1: Write pay off matrix.
Step 2 : Select the minimum value in each row of the pay off matrix under the head row
minimum written at right end of each row put a circle (0) over these elements
in the pay off matrix.
Step 3 : Select largest among the row minima found from step-2 and write beneath the
row minima.
Step 4 ; Select largest value (maximum) of each column and put a box or rectangle
(L_l ) over these values in the pay off matrix. Write these values of column
maximum beneath the pay off matrix.
Step 5 : Select minimum value of column maxima (found in step-4) and write this at the
right end.
Step 6 : If Max (R^) = Min (Cmax) i.e., value in step - 3 equals value in step-5 then
saddle point exits. Thus saddle point is found at the element on which both
circle and box are enclosed. Note this pay off element as the value of the game.
Game Theory 567

This is illustrated with the following numerical example.

ILLUSTRATION 1 :

Players Adithi andSahithi haue a competition to reach college early. Each has
two strategies viz going by car or going by a two-wheeler. If both go by cars,
Sahithi reaches 5 minutes early but if both go by two-wheelers, Adithi can reach
4 minutes early. If Adithi goes by car andSahithi by a two wheeler, Adithi reaches
2 minutes early while Adithi by two wheeler and Sahithi by car keeps Sahithi
7 minutes early. If both can choose their strategies independently and
simultaneously and know these estimated pay offs each other, find who will win.
What will be outcome of the game and their optimal choices.

Solution :
Step 1 : To Write Pay OffMatrix .Let us first formulate the information into the pay off
matrix.
Assume Adithi is the row player and Sahithi is column player. With this
assumption, we can say that gains to Adithi are positive and that to Sahithi are negative.
The pay offs for various alternatives are as follows.

Adithi by car when Sahithi by car is -5 (since Sahithi wins by 5 min).

Adithi by car when Sahithi by two wheeler is 2 (since Adithi wins by 2 min).

Adithi by two wheeler when Sahithi by car is -7 (since Sahithi wins by 7 min)

Adithi by two wheeler while Sahithi by two wheeler is 4 (since Adithi wins by 4
min).

Now, these pay offs are arranged in a matrix form as follows :

Sahithi
Car Two Wheeler
Adithi Car -5 2
Two Wheeler -7 4

Step 2 : Find row minima and encircle these values.

Sahithi
Car Two Wheeler J?min
Adithi Car o 2 -5
Two Wheeler Q 4 -7
568 Operations Research : Theory and Practice

Step 3 : Find maximum of row minima

Sahithi

Car Two Wheeler Rn

Adithi Car 2

Two Wheeler 0 4 -7
- 5 <- Max (Rmm)

Step 4 : Find column maxima and enrectangle these elements.

Sahithi

Car Two Wheeler jRmin

Adithi Car -5

Two Wheeler -7

- 5 <- Max (tfm

Step 5; Find Minimum of column maxima

Sahithi
Car Two Wheeler Rmin

Adithi Car y 2

Two Wheeler
y -7
_ gs- 5 <- Max (Rum) (Maximin)
O
Min(C max )
(Minimax)
As Maximin = Minimax or Max CRmin) = Min (C m a x ) the saddle point exists at
first row-first column element which is both encircled and enrectangled.

The Optimal Strategies :

For Adithi — TO GO BY CAR
For Sahithi — TO GO BY CAR
Value of the game = -5
Which means Sahixiii will win by 5 minutes or Adithi will lose by 5 minutes.
[It is a deterministic (pure strategies) unfair game]
Game Theory 569
ILLUSTRATION 2

Solve thefollowing game

Bj B2 B3 B4
-5 -2 0 7
A A2 5 -6 -4 8
A3 4 0 2 3
Solution :
Finding Rrr^n and then maximim among i?D

B2 £
4 min

A A,
0 -2 0 7 -5

5
Q -4 8 -6
4
® 2 3 0

Find C max and then min of C max

B% BA Rm\,
4 ^min

A 4i
0 -2 0 7 -5

5
0 -4 8 -6

4 K5) 2 3
0 Max(i? mill )

Min(C m a x )

Saddle point exists atA$, B%

A uses pure (deterministic) strategy A§

B uses pure (deterministric) strategy B<^
and value of the game (v) = 0
i,e., Game is drawn (i.e., neither^ nor B wins)
(This is a fair game as v = 0 )
570 Operations Research : Theory and Practice

11.4 Dominance Principle

The dominance principle is used to reduce the order of the pay off matrix.
Thus if saddle points exist, the matrix can be reduced to a matrix with one element
which is equal to value of the game.
Similarly, we can say that the dominance principle is used to delete rows and/
or columns of the pay off matrix which are inferior (i.e., less attractive) to at least one
of the remaining rows and / or columns (strategies) in terms the pay offs. The deleted
rows or columns would never be used by both the players for determining their
optimum strategy.
The dominance principle is composed of the following rules.
Rule 1 : Row Dominance :
For a row player, in a pay off matrix, if every element in a particular row is less
than or equal to the corresponding element of another row, then the former row is
said to be inferior or dominated by the latter. Therefore the row player will never
employ the former row (strategy) for whatever strategy that is used by nis opponent.
Hence this row can be deleted from the pay off matrix for further iteration.
B

i?2 dominates R\
since R%i>Ru (Ri
e.g. can be deleted)
R*

In the above pay off every element in R j is less than or equal to its corresponding
element in i?2- (4<5, 6=6 and -5<-3). Therefore Rl can be deleted since row player
(A) will never use R± for whatever strategy B uses among Cj, C2, and C3.
Note : The rule cannot be applied even if one element does not obey rule.

Rule - 2 : Column Dominance :

For a column player, in a pay off matrix if every element in a column is greater
than or equal to corresponding element of another column, the former column
strategy never yields any better result than the latter and therefore the former column
will never be used whatsoever may be the opponent's strategy. Hence the former
column can be deleted from the pay off matrix for further iteration.
B

2 C% dominates C\
e.g. since C%j< C\j
R* «iiii -1
(C\ can be deleted)

In the above pay off matrix, every element of Cj is either greater than or equal
its corresponding element of C2- This suggests* that, first column is inferior to the
second in any case and hence deleted.
Game Theory 571

Rule 3 : Modified Row Dominance :

When no single row (pure) strategy has dominance over the other, then the
comparison may be made between a row and the average of a group of rows, if every
element is less than or equal to average of corresponding elements of a group, then
the former row can be removed. If a row dominates over the average of group of rows,
then the group may be discarded.
Player B

e.g. 1 Player^

In Ax ScA2

Ax8cA3

A2 8c A%

Thus there is no pure row dominance.

But by comparing the Ax average values of A2 andA$ we have,

>2 of/^ and

>2

Therefore the row A^ is inferior to average of A% andv43. Hence A^ can be

"5 0]
deleted thus we reduce matrix to
0 8J
B

In the above pay off, between

Ax8cA2 8< 15 but 3 > - l
Ax8cAs 8>1 but 3<4

A2ScAs 15 > 1 but -1<4

Thus there is no pure dominance.
But comparing average of^42 &^s with^j, we have
572 Operations Research : Theory and Practice

8cAs = 8<8

-1+4 3
= 9"-

Therefore the average is dominated by^4x and hence both^42 and A% are deleted
leaving the matrix as [1 4]
Rule - 4 : Modified Column Dominance :

Similar to the above rule, when no pure column dominance exists, a convex
linear combination i.e., average of certain pure strategies may be compared with
another pure strategy applying rule 2, i.e., if every element in a column is greater than
or equal to corresponding element of other column the former is omitted.
Player B

8=
15+1
Player ,4 HI 15 1 2
e.g. -1 + 4
iHSl -i 4
Clearly, their is neither pure column nor pure row dominance exist in the above
pay off. But the average (convex linear combination) of 5 2 and B$ is greater than or
equal to corresponding element of first column. Therefore B^ can be deleted.
Thus we get,

P i]
These rules used to solve the games are illustrated through following numerical
examples.
ILLUSTRATION 3
Let us the take the game of illustration -1 and solve by using dominance :
After formulating we get
Sahithi
Car Two-wheeler
5
Adithi Car
Two-wheeler
isrffe C2j > = Cij (C% can be deleted)

Solution :
Now, from Sahithi's point of view, she never uses her second strategy i.e.,
two-wheeler since, for what ever Adithi's strategy, she will be the loser. Therefore her
car strategy dominates over two-wheeler strategy (every element of first column is less
Game Theory 573
than or equal to corresponding element of second column). Hence Sahithi's
two-wheeler strategy is deleted.

Thus we get,
Sahithi
Car
Adithi Car
R% <R\i (i?2 can be deleted)
Two-wheeler

Now, as Adithi has her better strategy with car (her loss is less here), she will
never got for two-wheeler strategy. (Every element in first row is greater than or equal
to the corresponding element in the second row). Hence second row is deleted.
Now the game is reduced to a single element matrix i.e., both using car strategy
with value of the game as - 5. i.e., Sahithi wins by 5 units.
ILLUSTRATION 4

An engineering student was frequently absent to the classes in a semester. To

safe guard himself he can choose one of the alternatives given below and the
professor also had four strategies. The student has approximated the probable
percent of marks in thefollowing pay off matrix against various strategies.
The students strategies are showing reasons as
Sx: due to ill health S2: to attend sister fs marriage are S3: went on project work S4
: attended inter college celebrations. The professor's strategies are Px: Not
giving attendance P2: Giving exam tough P3: Evaluating strictly PA:
Complaining to principal.
Thepay of is

1*2 1*4
55 53 32 62
s2 4O 30 74 5O
s3 57 54 44 53
54 54 72 56

Use dominance principle so that the student may choose his optimal strategy.

Solution :
In the given pay off, by examining thoroughly we find that every element of
P 2 column is less than or equal to the corresponding element of Pj column, hence
professor will never use the P± strategy (P2 dominates P^, hence Pj is deleted. The
pay off is then reduced to
574 Operations Research : Theory and Practice

P\j>P<2j {P% dominates and Pi is

deleted). After Pi is deleted S$ can be
deleted in the next iteration, Since

deleted
From the above pay off, it is clear that, the student will never use 5 S because
every element of 5 4 is greater than or equal to corresponding elemerit of 5 3 (5 4
dominates S3 ). Hence by using the principle of dominance we delete S3 and revise
the pay of as
p
t p P

Si. 53 • 32 ^ ^ ^ p P4j>P2j (P4 is deleted)

s2 30 74" ^ ^ ^ ^

s4 54

Again P 2 dominates P 4 , hence P 4 is deleted, we get

P
3
i (S\ is deleted)

s2 30 74
54 72

Now, S 4 dominates Sj, hence S± is deleted.

P3/ > P% (P3 can be deleted)

Now, As P 2 dominates P 3 , P 3 is deleted

(52 can be deleted)

54

Now, S4 dominates S2, hence S2 is deleted. Thus student uses S 4 strategy while
professor uses P 2 strategy i.e. The professor gives the exam tough when student claims
that his absence is due to attending intercoUege celebrations and the value of the game
is 54.
Game Theory 575

ILLUSTRATION 5

Let us consider the example in illustration - 2.

B
Bj
**1 BJ*2
2 B3 B4
Ai ' -5 -2 0 7
5 -6 -4 8
4 0 2 -3

S o Iution i
In this game, on observing we can notice the dominance of B2 over B$ and JB4.
(by Rule - 2)
The game reduces to
B

Now, A% dominates over Al and^ z is deleted, So, we get

B

-6

mm o
Now, as B% dominates B± and so, we delete
B

A$A$ is dominating A2 we have the value of game as 'zero'.

i.e., Fair game (drawn)
A uses Aj strategy and B uses B2 strategy.
576 Operations Research : Theory and Practice

Modified Dominance :
ILLUSTRATION 6 ——-
Use dominance principle to reduce thefollowing game to 2*2 game, Is the game
stable. nw n
Player B
Bt B2 B3 B4
6 -10 9 0
Player A
A2 6 7 8 1
A3 8 7 15 1
A4 3 4 -1 4
Solution ;

We can easily verify that the above game has no saddle points. Therefore it is
not stable.

Now using row dominance, the third row^43 dominates overy4x as well
Therefore^ and^42 are deleted (shaded in the pay off matrix of the problem given
above).
Thus the play reduces to
Player B
£>i JDc, £>o BA

Player^

Now, we can delete second column ( B2 strategy) as it is dominated by fourth

strategy (every element is less than or equal to corresponding element).
Player B
1 3
15
Player^
A, IKS 1

A4 SKtif -i 4

Now, there is no pure dominance, however we can apply modified dominance

with a convex combination of B's B$ and B± strategies dominates his first strategy.

=8<8
Game Theory 577
Therefore B's B±, strategy is deleted.
Thus the game reduces to
Player B

Player ,4

This game can be solved by using either algebraic or any other suitable method,
which will be explained in the sections to follow.

1L5 Games Without Saddle Points: Mixed Strategy

In certain cases, no pure strategy solutions exist for the game. In other words,
saddle points do riot exist. In all such game, both players may adopt an optimal blend
of the strategies called Mixed strategy to find a saddle (equilibrium) point. The
optimal mix for each player may be determined by assigning each strategy a
probability of it being chosen. Thus these mixed strategies are probabislitic
combinations of available better (or dominant) strategies and these games hence
called probabilistic games.
Summarily, when saddle points do no exist in the game, we can use the
dominance principle to reduce to 2 x 2 game or at least 2 x n or m x 2 matrices. If this
is also not possible we can formulate it as Linear programming problem.
Thus the probabilistic mixed strategy games without saddle points are
commonly solved by any of the following methods :
S.No. Method Applicable to
L Algebraic method 2 x 2 games
2. Graphical method 2 x 2 , m x 2 and 2xn games
3. Linear programming method 2 x 2 , ffix2, 2xn and m xn games.

These are explained here below :

1. Algebraic Method :
Suppose a general game without saddle points as given below :
Player B

player A

Since it is assumed that there are no saddle points, A wishes to use a mixed
strategy ofA i and A% with the probabilities of/? i and p%t and B to use a mixed strategy
of B i and JB2 with the probabilities of qy and q% respectively.
578 Operations Research : Theory and Practice
According to the assumption that A always tries to maximise his minimum
returns and B tries to minimise his maximum losses, they employ their strategies most
optimally such that the B's minimax losses will be equal A9s maximin gains.
Hence, out come of probability mix (of px andjf?2) by A in his Ax and^42
strategies when if use's his Bx strategy i.e., apx + cp2 is the value of the game.

Otherwise, when B uses his B2 strategy, A's returns will be apx + dp2, is also the
value of the game. These both must be equal.

(a-b)px = (d-c)p2

— =~ (i)
Pt a ~ b
Similarly, from B 's. point of view
aqx + bq2 = cq^ + dq2
i.e., (a-c)ql = (d-b)q2

q2 (a - c)

From the rule of probabilities, we know thatpi +p2 = 1 and qx + q2 = 1

Thus we have
or
P\ = l -P2 P2=l ~Pv • • • • (iiO
and qx = 1 - q2 or q2 = 1 - qx . . - . (iv)

Thus we have (on solving the equations )

(d-c) (c-d) (c-d)
P
-b) (c-d)-(a-b) (b + c)-(a + d
(d-b) = ib-d) _ (b-d)
b) + (a-c) (b-d)-(a-c) (b + c)-(a + d
Wherep2= l-px, q2= 1 -qx,
(and v - api + cp2 or bpx + dp2 or aqx + bq2 or c^j + d^2 )
ad - be
i.e., v=
(a + d) - (b + c)
This method is illustrated with the following numerical example.
Game Theory 579

* Illustration - 6 is continued :
Previous to this section, the problem in Illustration - 6 was reduced from
4 x 4 to 2 x 2 and left unsolved there.
Now we can use the above formulae and find the solution.
Player B
B
3

15 1
Player^
A4 -1 4
Assuming the probabilities p] 8cp2 for A and q± h!c£2 for B to use
t and 2*3, B4 respecitvely, we have

-1-4 14
H- 19
1-4 3 16
(-1 + 1)-(15+ 4) " 19 19
(15) ( 4 ) - ( - ! ) ( ! ) 61
( )( )

[ 5 14^
—, — while B uses his £3
w
h
. ., u _ ... . , u i, • , , . r u -61 . . .
strategies with probabilities —, — and the value of the game = — i.e., A wins.
r li/ 1 J I J. ZJ

ILtUSTRATIOK 7
Children Srija and Himaja play a game who have some 25^ paise coins and
SOpaise coins. Each draw a coin from their bags with out knowing other rs choice.
If the sum of coins drawn by both is even Srija wins them, otherwise Himaja wins.
Find the best strategyfor each player and alsofind the value of the game.
Solution :
First let us formulate the game. According to the data given Le.s each has some
25 paise coins and some 50 paise coins in their bags and draw one at a time.
Now, from this we can say that each has two strategies.
Strategy 1: Drawing 25 paise coins.
Strategy 2 : Drawing 50 paise coins.
Let us choose Srija as row players and Himaja as column player.
Now if Srija draws 25 paise (say strategy Sj) and Himaja draws 25 paise (say,
strategy H±), then the sum is even. Therefore Srija wins them. Her gains (pay off) is
25 paise.
Similarly, if Srija draws 25 paise (strategy S^) and Himaja draws 50 paise (say
strategy i/ 2 ) then Himaja wins these coins and her gain (pay off) is 25 paise or
otherwise Srija's loss i.e., -25 paise.
580 Operations Research : Theory and Practice

Again, if Srija draws 50 paise (strategy S 2 ) and Himaja draws 25 paise ( strategy
Hx), then Himaja wins 50 paise or Srija's loss i.e., - 50 paise.
Also, when both draw 50 paise ( S2 and H% ) Srija wins 50 paise.
Thus, the pay off is as follows.
Himaja
Hl (25 paise) H% (50 paise)
Srija Sx (25 paise) 25 -25
(50 paise) - 50 50
On examination we can find that there are no saddle points/ therefore the
players use their mixed strategies, say with the probabilities as fa andjf?2 for Srija on
Si and S2 respectively, and as q± and q% for Himaja on Hi andi/ 2 respectively.
Thus we have,
25px-50p2 =-25pi + 5Qp2 (1)
25qi-25-q2=-5Oql + 5Qq2 . . . . (2)
Also P1+P2 = 1 -. - • • (3)

From(l) 50^!= 100

r-fl • • • • (5)
Pz
From (3) and (5)
We get Pi=-; p2 = -
Also from (2) we have
75 qY = 7 5 ? 2

^=1 ,...(6)
n
From (4) and (6)

Thus Srija is expected to use her strategies 5 r andS 2 (i.e., Drawing

25 paise and 50 paise coins respectively) in probabilities of (2/3, 1/3).
While for Himaja it is optimal to use her strategies with (1/2, 1/2) probability
of H i and // 2 , i.e., drawing 25 paise and 50 paise coins respectively.
The value of the game is zero.
And it is fair game (V= 0)

[Since V - 25 x - | - 2 5 x | = 0

V = - 5 O x | + 5Ox|=

or 7=25xf-50x- = -
Game Theory 581
ILLUSTRATION 8

Find the best strategy and the value of thefollowing game:

I II III
I -1 -2 8
II 7 5 -I
III 6 0 12
[JNTU (CSE) 97/S]
ITERATION TABLEAU 1:

B
II III Row Minima
I -1 -2 8 -2
II 7 5 -1 -1
III 6 0 12 0 <- Max (Rmin)
Column 12 5
Maxima Min (CmJ

S o I M+ion \

By minimax (maximin) principle, we have no saddle points as

Max(# min )*Min(C max ).

Hence, we check the principle of dominance. It is found that Row ^4111 is

dominating AT, since every element in^4(III) is > the corresponding element oL4(I).

Therefore A will never use the strategy A(I), hence deleted. Now, the new
iterated pay off is
ITERATION TABLEAU 2 :

B
I II III
A II -1
III 12

The column 5(11) dominates B(l) since every element in 5(11) is less than its
corresponding element of B(I)..Hence B(I) is deleted (B will never use B-l as it yields
more losses against any strategy played byyl).
582 Operations Research : Theory and Practice

ITERATION TABLEAU 3 :
B
II III
II 5 -1
II 0 12
Now, they play a mixed strategies with probabilities. Since there are neither
saddle points nor any dominance.
Let A use the strategies II & III with probabilities^ and^ 2 respectively and B
use the strategies II 8c III with the probabilities q^ ancl^ respectively.
then Pi+p2= l>1l+<l2=1
And from the pay off we know that
and
p l
o A9l 13
— = 2 and — = -z~
5
Pl ft
2 1
13 5

. e 2 10
and the value of the game is 5 x —= —.
(by substituting in any of the above probability expressions).

Thus, the mixed strategies of A are (II, III) with the probabilities — and — .
L j
the mixed strategies of B are (II and III) with the probabilities (13/18 and 5/18).

The expected value of the game is — (A wins the game).

ILLUSTRATION 9
A and B play a game in which each has three coins a 5P, a 10P and 20P. Each
selects a coin with out the knowledge ofthe others choice. If the sum ofthe coins is
an odd amount, A winsBfs coins. If the sum is even B wins A fs coins. Find the best
strategy for each player and the value ofthe game.
[JNTU (Mech.) 99/CCC, (ECE) 99/CCC]

Solutiorv ;
When A draws 5p, if B draws bp, then B gains 5p of A, if B draws 10 p, A wins
\0p and if B draws 20^ then alsoy4 wins 10^ and so on.
Thus, the pay off matrix for player^ is formulated as :
Player B
5p 10p 20p

PlayeM bp
iOp
20p
Game Theory 583

It is clear that this game has no saddle point. Therefore, further we try to reduce
the size of the given pay off matrix. Note that every element in column £3 is more
than or equal to every corresponding element of column J52- Evidendy, the choice of
strategy B 3 by the player B will always result in more losses as compared to that of
selecting the strategy B%. Thus strategy B% is inferior toB%. Hence, deleting53 strategy
from the pay off matrix. The reduced pay off matrix is shown below:
Player B

Player^ A} -5 10
5 -10

After the column B§ is deleted, it may be noted that strategy A% of player^ is

dominated by his A$ strategy, since the profit due to strategy A% is greater than or
equal to the profit due to strategy A§, regardless of which strategy player B selects.
Hence strategy A% (row 3) can be deleted from further considerations. Thus the
reduced pay off matrix is
Player B

Player yl -5 10
5 -10
Since in the reduced 2 x2 matrix, the maximin. value is not equal to the
minimax value. Hence there is no saddle point and one cannot determine the point
of equilibrium. For this type of game situation, it is possible to obtain a solution by
applying the concept of mixed strategies.
Let p\ ,p2 and <?i » £2 be th e probabilities of A and B playing the strategies
{Ax ,A% ) and (Bl, B% ) respectively and we h a v e ^ +p2 = 1 and q\ + q%= 1.
Now, from the above pay off,
= 1 0 jf?a —

and - 5 q± + 10 q% - 5 q^ - 10 q%

Thus we have — = 1 and — = 2.

By solving the above equations, we have

l l 2 l
* * A

A n d t h e v a l u e of t h e g a m e is - 5 x — + 5 x — = 0 t h u s t h e m i x e d s t r a t e g y of

AQiiiA^) is w i t h t h e p r o b a b i l i t i e s o f —,— a n d t h a t o f B (Bx, B2 ) is — , — t h e

I J I J
value of the game is zero (fair game).
584 Operations Research : Theory and Practice

11.6 Graphical Solutions

Graphical solutions are helpful for the two-person zero-sum games of order
2 x n or m x 2 i.e., when one of the players have only two dominant strategies to mix
while other has many strategies to play. However, graphical solutions used with an
assumption that optimal strategies for both the players assign non-zero probabilities
to the same number of pure strategies. Therefore, if one player has only two strategies,
the other also should have to use same (two) number of strategies. This method is
helpful in finding out which two strategies can be used.

Here we have two cases viz 2.x n and m x 2 cases.

Case (i) : 2 x n games :

In this case, row player has two. strategies to play while column player has n
strategies. The graphical method is applied to find which two/strategies of column
player are to be used.
Suppose a pay off given below :
PlayerB
B
2 Bs • •:• Bn.

Player^ ^1 a
A2 a
22
a
23 a
2n

If player A mixes his strategies^ and A% with probabilities p± and P2-Q where
Pi + p2 = 1 » t n e n f° r e a c n strategy of B, A will have the expected pay off as follows.
Es pure- strategy A's pay off

B2

B
n \ 2
As A is assumed to always try to maximise his minimum gains. The highest point
of lower envelop (lower boundary) formed by drawing these as straight lines represents
the optimal probability mix: The lines corresponding to this point will yield a 2 x 2
matrix from which the value of the game, optimal strategies and their probability mix
can be found.
Thus the two series are to represented on two parallel axes and these are joined
to represent respective pay offs. Then the lower envelop is identified and upon this
highest point is located. The coefficients of these lines make a 2 x 2 matrix for further
iteration / calculation.
The above method is illustrated through the following example :
Game Theory 585
ILLUSTRATION 1 0 —
Solve thefollowing game graphically.
V-e o 6 -3/2
\_7 -3 -8 [JNTU (Mech.) 2001]
Solution :
Let Row player's (say^4) strategies are^j and^42 are used with the probabilities
p j and p2, then his expected pay offs when his opponent (column player) uses his pure
strategies are shown below :
Column players pure strategies Row player (A's) expected pay off

III
IV
These are graphically represented as follows as two parallel axes of unit distance
apart.
Ax s 2 Axi s i

9- •9

Q Q
o • • o

7- • 7

6- •6
\ /
/
•

5- • 5
\ I: -6p, + 7 p 2 /
4- \ •4

3- •3

2- •2
1- • 1

-1- •-1

-2- •-2

3- y\ \ Maximum \ \ / ^ , •-3
S
-4- (<W?ot \W^
-5- • / < <
•-5

-6- •-6
v •
-7-
-8-
/ \ \\\ N
•-7

•-8
Lower envelope \ \

FIGURE I 1.2 :
586 Operations Research : Theory and Practice
The highest point on lower envelop appear at the intersection of the lines
represented by column players I and II strategies.
7p2 and
The required 2 x 2 matrix is
?l ?2

- 3°1J
Pi r-6
Pi L 7
We have Pl + 7pl =* 6p ! = i0^ 2
P1+P2 = 1
10 5 3 3
Pi " 16 ~ 8 ' 6~8
and l-Sft =» 131?l = 3^2
_ 3
and ?1 "*" ?2 = ^[
__3_ 13
9.1 ~ 16 " 16
.\ Row player will use his strategies with the probabilities as (5/8,3/8)
respectively where the column player uses the first two strategies mix at the
probabilities of (3/16, 13/16).
The value of the game

or
The column player wins the game with a gain of 9/8 units or row player loses
the game with a loss of 9/8 units.

? (ii) : m x 2 games :•
In this case a column player has two strategies while row player can choose
optimal mix of best two among m strategies. The graphical method is usedto Find
which two strategies of rows player would be the best.
Suppose a pay off of two players A and B is given below
Player B
B B
l 2
u
an

Player A

a
m\
Game Theory 587
When player B mixes his strategies Bx and £ 2 with non-zero probabilities
and #2 where q± + q% = 1, then for each strategy of A, B's expected pay off is given
by
A's pure strategy B's pay off

a
\2(i\+a22<l2

Now, as B is assumed to be the looser always, he will try to minimise his

maximum losses. Thus when these expected pay offs of B, when drawn as straight
lines between two parallel axes the lowest point (minimising) of the upper envelop
(maximum losses) represents his optimal strategy. The line consisting of this point
will represent the best two strategies at its edges on the two parallel axes of unit
distance apart. TTie perpendicular drawn from the minimum point of envelop on to
x-axis divides segment on x-axis into two parts equal to the probabilities used by B.
This is illustrated through the following example.
{ I L L U S T R A T I O N 11
Solve thefollowing game graphically.

1 —3
3 5
-1 6
4 1
2 2
-5 O

I Solution ;
Let B's strategies By and 252 t*e used with the probabilities q^ arid# 2 where
= 1. Then the pay offs of B for each of the strategies used by A are given below.
A's pure strategy B's expected pay off
A
\ 9\~ ^?2
588 Operations Research : Theory and Practice

The above pay offs against A'5 strategies are drawn as straight lines between two
parallel axes of unit distance apart.
Since.column player B wishes to minimise his maximum expected pay off (loss),
we consider the lowest point (minimum) of intersection G in the upper envelop
EFGH (maximum) which represents the minimax of B's expected pay off.
The minimax point G is the point of intersection of the lines A2 and^44. Thus
the pay off ofA is found most optimal with the blend ofA% and^44 strategies. Thus the
6 x 2 matrix reduces to 2 x 2 matrix with A's A2 and^44 and B's Z^ and B% strategies
which can be evaluated by algebraic method.
Axis 1 Axis 2

FIGURE
B
«1 -
.5
1
Now, ifpi and^?2 a r e th e probabilities of^4 to and q± and q% are
the probabilities of B to play B x and B2 then
Game Theory 589

3p j'+ 4p2 = bpl+p2 = V (the value of the game)

and Sqy + bq2 = 4q\+ q2 = V (the value of the game)
Where pi +pi = 1 and "£i + ft = l-
Solving the above equations we get,

and pi + p 2 - 1
3 and 2
P\ " 5 "5
4
Also
$2
and ai + ao
4 ' 1
?l = "g" . a n d ft = 5"
plays his^2 and^44 strategies with the probabilities (3/5, 2/5) while 5
applies his B^ and JB2 strategies with the probability mix (4/5, 1/5) and the value of
the game
1T '4 1 17

Remark : Observe the strategies A j, ^5 an6? ^4g an^f compare each with A%- You can find
is dominating these three since it is up over these in graph

11.7 Linear Programming Method

When a two person zero sum games are in larger size (say mxn) these can be
solved by linear programming method. This is the major advantage of this method
and hence can be considered as universal method.
To illustrate this method first let us consider a 3 x 3 matrix given below :
B
*\ B2
Ax an an
A A2 a
21 Ht
A, a31 aS2

Now, as per the assumptions, A always attempts to choose the set of strategies
with the non-zero probabilities say p^p^^ndp^ ( where p\+p2+p3= 1) t n a t
maximises his minimum expected gains.
Similarly, according to B, he would choose the set of strategies with the non-zero
probabilities, say qlf q2 and q$. (where qY + q2 + q$ = 1), that minimises his maximum
expected losses.
From the above statements, we can draw the following conclusions.
590 Operations Research : Theory and Practice

From A!s. point of view ,

He would mix all his strategies A yAtfand A§ with non*zero probabilities
P\,p2?ti&p3 i n s u c n a waY t n a t t n e sum should be atleast equal to the value of the
game (V) for whatever strategy that B may play. In other words, the sum of his expected
pay off should be greater than or equal to V.
Therefore,
As expected gain is a>\\p\+ a>%\ p% + a 3i ^3 ^ ^when B plays By
Similarly, As gains when B uses his second strategy (B%) and third strategy
CB3) respectively are

and
And also px +p%+p$= 1.
Simplifying the above inequations, by dividing by V through out,

(Pi) (Pt) (Ps)

l j + l j +
pj.

Pi) (Pi) (Pi

and
, P\ p2p$^Ci
"F'^T" 0
Since, the objective of /I is to maximise the value of the game V, which is
equivalent to minimising the value of —, the objective function of^4 is

. . . 1 P2 P%
+
mimise A=y == V V+ V
Rewriting the above the assuming
Pi P% Pz
Xl — = x2 and — = x$,\
v '
The LPP formulation as
Minimise 2.£ —x^ + x2 + ^3
Subject to a
\ j xi + « 2 1 X2 + a
3 1 X2> -
av 2*1 + a
2 2 ^2 + ^32 X 3 -
u
l.
and > 0, x2 > 0, x 3 > 0
Game Theory 591

Similarly, from B '$ view, we get the L.P problem as the dual of the above
problem.

(Player B minimises his maximum losses).

For player B, it is written as
Maximise ZB- y\+y2+y$
+fl +a
Subject to a.nyi i2)2 l3 y$- *

and

where 3^ = y , y2 = y a n d j 3 == —

- This problem now can be solved using usual simplex method, and the problem
can be extended to any size mxn. Now let us illustrate this method with an
example.
ILLUSTRATION 12
Two companies A and B are competing for the same product Their different
strategies are given in thefollowingpay off matrix.
Company A
Aj A2 A3
Bj |~ 2 -2 3 "I
Company B
B2 ^-3 5 -1 \
Use linearprogramming to determine the best strategies for both the players.
[JNTU-CSIT-99]

Solution j
Assume the non-zero probabilities p\,p% for B to use his strategies By and B2
respectively ( where py +p2 = 1 ) and qy, q2, q<$ for A to use A^fA2fA^ respectively
(where qy + q2 + q$ = 1 )

Now with reference to A, Max ZA= —or — + — + —-

q2,
592 Operations Research : Theory and Practice
Diving by V through out the inequalities and equation,

# # # ' • -##•
(Jl $2 # 3 - 1
and —+—+—=—

Since A wishes to minimise his maximum losses F o r maximises — the objective

function of A will be
Maximise
Subject to 2xy- 2%2 + 3x 3 < 1
-3JCJ + 5 x 2 - xs< 1

and Jti,.*2» x 3 > 0

Converting the inequalities into equalities by introducing slack variables, we

get,
Maximise = x^ + x% + x$ + os-^ +
Subject to 2^i - 2x 2 + 3x 3 + 5^

- x% + $2=

lf52, >0
IBFS:
X| = 0, x 2 s= 0, X3 = 0, (non basic)
5
1 " 1» 5 2 = 1 (b a s i c )
ITERATION TABLEAU 1 :
- Entering variable

CB BV i 1 0 0 Min Remarks
Ratio
•SK
Wi x2 xs s2

0 7 1 5 -1 0 1 -ve (Ignore -ve ratio)

;>.<>• : ;;r

*0 0 0 0

Eeaving variable -1 -1 0 0

Key colurrm
[Here there is a tie for entering variables x^x^ and x§, all three being decision
variables, selection is made arbitrarily].
Game Theory 593
ITERATION TABLEAU 2 :

—-—-"
Entenng variable

BV JC
iii i 0 0 Min
Ratio
Remarks
SV

1 1/2 1/2 0 -ve

SP HHH 1 •' j j Hi
j
pi
; Z
J 3/2 1/2 0

Leavin
J g variable
z c
rj .0 .jjjHi 1/2 1/2 0

Key column
ITERATION TABLEAU 3 :
1 1 1 0 0 Min Remarks
Ratio
52
1 7/4 1 0 13/4 5/4 1/2 Af-»«? + / #
1 «2 5/4 0 1 7/4 3/4 1/2 i

Z; 1 1 5 2 1
z --
Z 0 0 4 2 1
J-fj
Since Z- - Cy > 0 for all the variables we have attained optimal solution.
The solution is

-- »- -0
4 4'
and Maximise
•4 4 4

•^ = 3 => value of the game (V) = —

and

"3 3

To find B's probabilities of using B x and B% we use Z - C values of final table

[by duality theorems].
594 Operations Research : Theory and Practice

Z':- C'j = 2 corresponds toy} = 2

and Zj - Cj for J 2 = 1 corresponds to j 2 = 1

Where 3?^, /y2 a r e the variables of dual problem of the above (primal) problem,

i.e., Minimise ZB = jya +^ 2

Subject to 2^! - 3))2> 1,

- 2 ^ + 5,2*1,

and

Where yx =
Thus

Thus^4 uses his^41? v42 strategies with probability mix as (7/12, 5/12).
While B uses his B Y, B2 strategies with probability mix as (2/3, 1/3) and the value
of the game= 1/3.
i.e., B wins by 1/3 units.

11.8 Distinction Between Deterministic and Probabilistic

Games
(Distinction Between Games With Saddle Points and Games Without Saddle Points)

S.No. Games With Saddle Points (Deterministic) Games Without Saddle Points (Probabilistic)
1. These games have saddle points. Saddle points do not exist in these games.
2. Pure strategies are used by the players in these Players use mixed strategies in these games.
games.
3. . The game is strictly deterministic. The game uses probabilities
4. These games can be solved by Mirii-max and Algebraic or linear programming techniques are
Maxi-min criteria or by dominance rules. used to solve these games. However, dominance
rules, graphical methods can be used to reduce
to 2 x 2 size and then -algebraic method is
applied.
Game Theory 595

11.9 Limitations .of Game Theory

The game theory was received with a great enthusiasm initially but later has
been found with lot of limitations. The most prominent limitations are listed here
below.

1. Unrealistic Assumption : The assumption that the players know their own as
well as other's (opponent's) payoffs in unrealistic. One can only guess of his own
and opponent's strategies or pay offs.

2. Complexity with More Number of Players : As the number of players increase,

the problem becomes so complex that is becomes difficult to solve.

3. Oligopolic Situations can not be Solved : Duopolic situations are easy to solve
but oligopolic situations can not be solved in many occasions by using Game
theory and in general practice, we have mostly, oligoplie situations in business.

4. Minimax - Maximin in Impractical: The assumption of maximin - minimax

show that the players are risk-averse and have thorough knowledge of
strategies, which seems impractical.

5. Limitation on Collusion : the game theory can not be applicable in case the
players share their business, secrets to workout a collusion.

6. Non-Dynamic or Static Nature : In markets, generally the changes are very

frequent and fast. The game thqory gives least scope to workout the strategy
immediately unless some ground work to known pay-off is made.

7. No base for calculating pay off : There is no correct formula and scientific
base for calculating pay offs of game.

11.10 Flow Chart of Game Theory

The theory of games and various solution methods applied are summarized
through a flow chart given in page 11-2.
596 Operations Research : Theory and Practice

Practice Problems
i GAMES WITH SADDLE POINTS
1. Solve the game when.pay off matrix is given by
Player B

Player A -3 12 -1
-6 -4 -3
-5 15 -8
[JNTU (CSE) 97]
Answer : Opt. Strategies (4\, By} with V= - 3 i.e., B wins by 3 units.
2. Verify your result in the above problem by dominance principle.
3. The pay ofFniatrix in respect of a two person zero-sum game is
As Strategy B's Strategy
Bx B2 *3 . *4
8 10 -3 -8. - 12
A2 3 6 0 6 12
7 5 -2 -8 17
A*. -11 12 -10 10 20
-7 0 0 6 2
(a) Write the maximin and minimax strategy.
(b) Is it a strictly determinable game?
(c) What is the value of the game?
(cl) Verify your answer using dominance principles. (CSE) 2000]

Answer; (a) Minimax of B with £3 - zero, Maximin of A with^42 - zero.

(b) Yes, it is strictly determinable game, since saddle point exists at (A%,
i.e., zero.
(c) Value of the game is zero i.e., fair game.
4. Consider the game with the following pay off
player B
B.i B2
A2\ 2 6]
player^ - ^ ^

(a) Show that the game is strictly determinable, whatever X may be.
(b) Determine the value of the game.
Answer: Value of the game is 2 with optimal strategy for player A and £ is
d j , respectively.
Game Theory 597
5. A company management and the labour union are negotiating a new three year
settlement. Each of these has four strategies.
I : Hard and aggressive bargaining.
II : Reasoning and logical.
III : Legalistic strategy.
IV : Conciliatory approach.
The costs to the company are given for every pair of strategy choice.
Company strategies
_ I II III IV
Union strategies I 20 15 12 35
II 25 14 8 10
III 40 2 10 5
IV -5 4 11 0

What strategy will the two sides adopt? Also determine the value of the game.
Use minimax - maximin rule and then verify your result with dominance rule.
[OU (MBA) 87]

Answer: Value of the game = 1 2 for which company will always adopt strategy - III
- legalistic strategy while union will always adopt strategy - I hard arid
aggressive bargaining.

6. Two competitive manufacturers are producing a new toy under license from a
patent holder. In order to meet the demand they have the option of running
the plant for 8, 16 or 24 hours a day. As the length of production increases so
does the cost. One of the manufacturers, say A, has setup the matrix given below,
in which he estimates the percentage of the market that he could capture and
maintain the different product schedules.
Manufacturer/! Manufacturer B
CA : 8 hrs C2 : 16 hrs C3 : 24 hrs
Sx : 8 hrs. 60% 56% 34%
5 2 : 16 hrs 63% 60% 55%
53:24hrs 83% 72% 60%.

(i) At which level should each produce,

(ii) What percentage of the market will B have. [JNTU (FDH, Med\.) 93/S]

Answer; (i) Optimal strategies : (S3, C3) i.e., both should produce 24 hrs a day
(ii) B loses 60% market share which is most optimum, otherwise he losses more
than 60%. Thus B will have 40% of market.
598 Operations Research : Theory and Practice

7. Two firms are competing for market share for a certain product. Each firm is
considering what promotional strategy to employ for coming period. Assume
the following pay off matrix describes the increase in market share of the firm
A and decrease the market share for firm B. Determine the optimal strategies
for each firm.
Firm B
No promotion Moderate promotion Much promotion
No promotion 5 0 -10
Firm A Moderate promotion 10 6 2
Much promotion 20 15 10

(i) Which firm would be the winner in terms of market share?

(ii) Would the solution strategies necessarily maximise profits for either of the
firms.
(iii) What might the two firms do to maximise their profits?
[JNTU (Hech. & ECE) 94/S]

Answer: Optimal strategy for both^4 and B is "Much promotion" with value of game
= 10.

(i) Firm A is the winner.

(ii) Yes, because it is a zero sum game,
(iii) They have to go for much promotion.

8. Solve the game using dominance principle and verify with saddle points of
minimax - maximin.
Firm B
B2
3 -1 4 6 7

A2 -1 8 2 4 12
Firm ,4
*s 16 8 6 14 12
A, 1 11 -4 2 1
[IGNOU (Assignment/MBA) 1996]

Answer: Optimal strategy is (A$, Bs) with value of game = 6.

Game Theory 599
9. Solve the following game by minimax - maximin and then verify by dominance
rules.
9 3 1 8 0
6 5 4 6 ,7
2 4 3 3 8 *
5 6 2 2 1
[IGNOU (MBA) 97]
Answer : 4 in second row, third column.
10. Solve by dominance and verify by finding saddle points.
—2 0 •o -5 3
3 2 1 2
-3 0 . -2" 6
5 3 -4 2 -6
[Dr. BRAOU (Assignment) 97]

Answer : Value = 1 in second row and third column.

11. For what value of X, the game with following pay off is stricdy determinable.
X 6 2
-1 X -7
-2 4 X
[Dr. BRAOU (Assignment) 97]
Answer : - 7 > X > 6.

12. Solve the following games by minimax - maximin

[JNTU (Mech., FDH) 96]

Answer: Value with the game = 2 when row player uses his 2nd strategy while
column player can use either 3rd or 4th strategies. Two saddle points.
13. Solve the following games by dominance.
3 -5 0 6
-4 -2 1 2
5 4 2 3
[JNTU (Mech., FDH) 93]

Answer : Value = 2 for 3rd row and 3rd column.

600 Operations Research : Theory and Practice

14. Use rrrinimax - maximin and verify by any other method to solve the following
game.
I II III
.8 6
II 12 2
Answer: V= 6 with (I, II)
15. Solve the game and check your answer by graph.
-1 -2
6 -6
Answer: (I, III), V= -2
16. Determine the range of value of p and q that will make the pay off element
a22 a saddle point for the game whose payoff matrix (a^) is given below :
Player B
4 5 [JNTU Mech/Prod/Chem/
Player A 10 .7 q Mechatronics 20G1/S]

17. Use dominance principles to solve the following.

5 1 10
50 1 1
50 0.1 10
Answer: (A%, B2), V= 1 [JNTU Mech. 2002/C/P]

III GAMES WITHOUT SADDLE POINTS-

1. Solve the following game without saddle points.
Player B
[JNTU (CSE) 2001]
Player^

29 (4 3^\ (2 5^
Answer : A wins with —, A uses probl. —, — and B uses with prob hr/"^"
2. Use concept of dominance to reduce the size of the matrix of given problem to
2 x 3 matrix and solve the game.
Player B
1 8 3
PlayeM 6 4 5
[JNTU (CSIT) 97/S]
0 1 2
3. Consider the game
B
5 50 50
1 1 0.1
10 1 10
Verify that the strategies (1/6, 0, 5/6) for player A and (49/54, 5/54,0) for B are
optimal and find the value of the game. [JNTU (CSE) 2OO3/S (Set - 1 ) ]
Game Theory 601
4. Use dominance principles to solve the following.
-2
6
[0U (MSc.) 89]
5. By modified dominance method solve the following
4 4 2 -4 -6
8 6 8. -4 0
10 2 4 .' 10 12
[OU (MSc.) 92]

Answer ; Row player (II, III) : - , - ; col player (II, IV) :•[-, j - :
iy y i iy y ~9~
V
6. By L.P.P solve the following game
3 -1 -3
-3 3 —1
-4 -31 3

/ 2 0 11 14^ 11 20^|
Answer
45'45'45 } <#• 45' 45
7. By L.P.P solve the following game
45

—1 2 - 1
1 -2 2
3 -4 -3

20_n
46' 46 I' ^ [23'23'23 23
8. By L.P.P solve the following game
7 • 1 7
9 -1 1
5 7 6

28
B
[3.0' 30' 30 5
9. By L.P.iP solve the following game

9 I 4
0 6.1 3
5 2
00.

Answer :•
• [h
602 Operations Research : Theory and Practice

10. Solve the game whose pay off matrix,

3 2 4 0
2 4 3 4
4 2 4 0
0 4 0 8

11. Explain the notion of dominance and use it to solve the game.
0 0 0 0 Q 0
4 2 0 2 1 1
4 3 1 3 2- 2
4 3 7 -5 1 2
4 3 4 -1 2 2
4 3 3 -2 2 2

Answer: Maximising player i 0, 0, —, -z9 0, 0

Minimising player J10, 0, — -z, 0, 0 L and

12. Solve the following two person zero-sum game.

Player B

5 7 11
Player ,4 2 -1 8
18 -6 10

Find the optimal strategies for the two players and the value of the game to
each of the player. [JNTU (Mech.) 98/S]
Answer: Player/I uses (Ax,As ) - ( 12/13, 1/13)
Player B uses (Blt B2):( 1/2, 1/2) value of game = 55/26 i.e., A wins.

13. Solve the following games graphically.

2 5 -2. 6
-2 -3 1 0

Answer : V= \ - | j f|, | 1 and | | , 0, 0, | , 0

Game Theory 603

14. Solve the following games graphically.

is ( 2 n (2 i
Answer: V = T , A \ 0 , - - \ B \ - -

15. Solve the following games graphically.

la) 2 7 (b) 1 3 11 i
3 5 8 5 2
11 2,

) 5
^4nsu/er : JjB
4' '14 l ll4' HI
16. Solve the following games graphically.
6 2 7
1 9 3

T/ 13 f2 n r r 5
Answer: F a a - - - - -
17. Solve the following games graphically.
6 8 6
4 12 2

nswer: V= 6, (I, III)

18. Solve the following games graphically.

-1 2 -2
6 .4 -6

Is the game (i) fair (ii) strictly determinable?

Answer : Yes, Strictly determinable. V= - 2 of 1st row, 3rd column
19. Solve the following games graphically.
6
-1

Answer : V= | fe | ] fo ±, | , ol or |0, ?, 0, i |.
604 Operations Research : Theory and Practice

20. Solve the following games graphically.

21. Determine the best strategy for player^ in the following game :
Player B

I II
I 3 -5
II 1 -1
Player^ III . 2 -3 .
IV -1 3
V 0 1

Answer:' In case of graphical solution all the lines pass through the minimax point,

{ 1 2 1 fl 81 T 2 1 1

- , o , o , - , o l , J-,o, o, o,-L Jo, —, o, —, o ,

) , 0, —, 0 , - 1 value of the game is —.
DO O

22. Solve the following game graphically.

1 -1. 2 -5
-3 6. 2- 0
[JNTU (Mech.) 94/CCC]

23. Solve the followng game graphically

* ] • B2 *3

2 . -4 6 -3 5

-3 4 -4 1 0

14 5l f4 5 1 7
Answer: A - , - L B - , 0. 0, -, 0 , V=--
Game Theory 605

24. Obtain the optimal strategies for both players and the value of the game for
two-person zero-sum game whose payoffs matrix is given as follows :

Player B

Player^

Answer: A (AVA2); —, — ; B (*„ B2): ^ - , — ; V= —

25. Use graphical method to solve the game

PlayerB

-6 7
A, 4 -5
PlayerA
-1 -2

A4 -2 5

7 -6

[ICWA, June 1985]

26. A soft drink company calculated the market share of two products against its
major competitor having three products and found out the impact of additional
advertisement in any one of its products against the order.

Company/I

What is the best strategy for the company as well as the competitor? What is the
pay off obtained by the company and the competitor in the long run? Use
graphical method to obtain the solution. [Delhi. Univ., MBA April 1989]
606 Operations Research : Theory and Practice

27. Consider the following payoff matrix for two firms. Find the best strategies for
both the firms.

Firm II

No advertising Medium advertising Large advertising

No advertising 60 50 40
m
* Medium advertising 70 70 50
Large advertising 80 60 75

3 4
Answer ; Firm I:\0,—,—

28. Two children Divya and Savya play the following game, named "scissors, paper,
stone". Both players simultaneously call one of the three "scissors, paper and
stone" scissors beat paper (as paper can be cut by scissors), paper beats stone
(as stone can be wrapped in paper) and stone beats scrissors. If the two players
name the same item, then there is a tie. If there is one point for win, zero for
tie and -1 for loss, form the pay off matrix of the game and solve it.

Answer : The game is symmetric. Hence V- 0 optimal strategies for both players
r
l 1 H

Hint : Savya
Sc P St

Sc 0 1 -1
Divya p -1 0 1

St 1 -1 0

29. Solve the following games by LPP.

1 -1 3
3 5 -3
6 2 -2
[JNTU (Mech., FDH) 94]

Answer: V- 1, for maximising player J —, —, Oj for minimising player \ 0, —, — L

Game Theory 607

30. Two firms A and B are competing for the same product. Their different
strategies are given in the following pay off matrix.
Company A

Company B

Use linear programming to determine the best strategies for both the players.
[IGNOU (MCA, Assignment) 96]

(1 <5
Answer : A (Ah A2) : —, —

31. Raju and Ravi plays game as follows they simultaneously and independently
write one of the three numbers 1, 2 and 3. If the sum of the numbers written is
even, Ravi pays to Raju this sum in rupees. If it is odd, Raju pays the sum to
Ravi in rupees. Form the payoff matrix of player Raju and solve the game to
find out the value of the game and probabilities of mixed strategies of Raju and
Ravi. [DOECCA 98]

2 - 3 4
Answer: -3 4 -5

32. For the following pay off matrix^ find the value of the game and the strategies
of players A and B using linear programming.

-1 —2
B
6 -6
[JNTU (Mech., FDH) 92]

Answer ; i, A<$), v = - 2; Game is deterministic and unfair

608 Operations Research : Theory and Practice

33. Find the best strategy and the value of the following game, using LPP method
B
I II III
I -1 -2 8
A II 7 5 -1
III 6 0 12
[JNTU (CSE) 97/S]

Answer : A uses (II, III) with prob. (2/3, 1/3)

B uses (II, III) with prob. (13/18, 5/18) value of the game = 10/3
i.e., A wins.

34. A and B play a game in which each has three coins a 5 paise, a 10 paise and a
20 paise. Each selects a coin without the knowledge of the others choice. If the
sum of the coins is an odd amount, A wins B's coins. If the sum is even B wins
A's coins. Find the best strategy for each player and the value of the game. Use
LPP method to solve the game. [JNTU (Mech.) 99, (ECE) 99]

Answer: Player^ (AlfA2) : (1/2. 1/2),

Player B CB1; B%) : ( 2/3, 1/3) value of the game is zero i.e., fair game.

^3
-5 10 20'
Hint: nx
5 -10 -10
pay-ofFsA A2
5 -20 -20
A*
35. Firm X is fighting for its life against the determination of firm Y to drive it out
of the industry. Firm X has the choice of increasing price, leaving it unchanged
or lowering it. Firm Y has the same three options. Firm Xs gross sales in the
event of each of the pairs of choices are shown below.

Firm Ys pricing strategies

Increase Do not Reduce
price change price
Increase price 90 80 110"
Firm X's pricing strageties Do not change 110 100 90
Reduce price 120 70 80

Assuming firm X as the maximising one, formulate and solve the problem as a
linear programming problem. [Osmania MBA, Nov. 1990]
Game Theory 609

Review Questions
1. Explain (i) strategy (ii) pay off matrix (iii) saddle points.

2. What is optimal strategy? Discuss rectangular games without saddle points.

3. Write short notes on solution of games with saddle points.

4. Write short notes on

(i) Applications of game theory in business systems.

(ii) Distinguish between games with saddle points and without saddle points.

[3NTU Mech/Mechatronics/Chem/Prod. 2001/S

5. Write short notes on minimax criteria. [JNTU Mech. 1998/S

6. Distinguish between zero-sum and non-zero sum game.

7. Define a rectangular game. Briefout its solution methods.

8. Distinguish between deterministic and probabilistic games. What approaches

do you apply to solve each of them.

9. Explain the approaches to solve rectangular games through a flow chart.

10. How is concept of dominance used in simplifying rectangular games.

11. Explain the graphical method of solving 2 x n games and m x 2 games.

12. How do you apply game theory in the following fields.

(a) marketing (b) personnel department.

13. Explain the relevance of game theory to managerial problems.

14. Explain how you can apply linear programming to game theory.

15. Give a critical appreciation of game theory with reference to duality concept in
LPP.

16. Write short notes on

(a) Two-person zero-sum games.
(b) Pure and mixed strategies.
610 Operations Research : Theory and Practice

17. Distinguish the following

(a) Two-person Vs. multi-person games.
(b) Zero-sum Vs. non-zero-sum games.
(c) Fair Vs. unfair games.
(d) Deterministic Vs. probabilistic games.
18. Distinguish the terms minimax and maxi-min with reference to rectangular
games. [JNTU Mech/Prod/Chem. 2001/S]

19. Discuss briefly different types of rectangular games and their solution methods.

20. Give the algorithm of finding saddle points in a rectangular game.

21. List put the assumptions made in the theory of game.
[JNTU CSE/ECE 2001]

Objective Type Questions

1. If the value of the game is zero, then that game/strategy is known as
(a) pure strategy (b) mixed strategy
(c) fair game (d) pure game

2. When minimax and maximin criteria match, then

(a) saddle points exist (b) fair game is resulted
(c) mixed strategies are adopted (d) game will be unfair
3. A predetermined plan of action based on which the games are played that does
not change during the game is said to be
(a) fair strategy (b) pure strategy
(c) mixed strategy (d) value of the game

4. Which of the following is true in the case of row dominance in a game theory
(a) least of the row > highest of another row
(b) least of the row < highest of another row
(c) every element of a row > corresponding element of another row
(d) every element of a row < corresponding element of another row
5. If the gain of a player is loss of another player then the game is called
(a) fair game (b) unfair game
(c) zero sum game (d) non zero sum game
Game Theory 611

6. Which of the following is never called a two person game

(a) chess (b) carrons
(c) Horse race (d) cricket

7. In a two person-zero sum game, the following assumption is wrong.

(a) row player is always a loser
(b) column player always minimises losses
(c) there are only two persons
(d) if one loses, the other gains
8. Games with saddle points are in nature
(a) deterministic (b) probabilistic
(c) stochastic (d) normative

9. If a two person zero sum game is converted to LP problem

(a) no. of variables are two only
(b) no. of constraints are two only
(c) if row player represents primal, column player represents dual
(d) it will not have objective function

10. Which of the following wrong

(a) games without saddle points are probabilistic

(b) games with saddle points will have pure strategies
(c) games without saddle points use mixed strategies
(d) games with saddle points can not be solved by dominance rule

11. If saddle point does not exist,

(a) it is deterministic game

(b) mixed strategies are used
(c) maximin - minimax criteria are used
(d) it is a fair game

12. A game is said to be fair if

(a) saddle point exists (b) pure strategies are used

(c) value of the game is zero (d) it is a zero sum game
612 Operations Research : Theory and Practice

13. Which of the following method can not be used for the games without saddle
points

(a) algebraic method (b) linear programming

(c) graphical solution (d) minimax - maximin criteria

14. If there are more than two players, the game is

(a) non-zero sum game (b) multiplayer game

(c) zero sum game (d) open game

15. If loss in sales of pepsi is gain in sales of coke, then the game between pepsi
and coke is a

(a) fair game (b) pure game

(c) zero - sum game (d) one - one game

16. In a colony there are 100 families which either use Colgate or pepsodent tooth
paste. Due to raise in price of Colgate some of them discontinued to use any
tooth paste. This situation can be fit in

(a) zero sum game (b) positive sum game

(c) negative sum game (d) fair game

17. The value of a game whose pay off is a 2 x 2 unit matrix is

(a) zero (b) unity

(c) 0.5 (d) none of these
18. Which of the following is a fair game

[-
t
1 C
(b)
0
1
(d)
1
"I]
r i
3
21
19. The value of the game whose pay off is
[-
(b) 2
2
J
(c)-2 (d) 3
Game Theory 613

-4 6
20. Value of the game whose pay off is given by -3 - 2 will be
-5 4

(a) zero (b)-3

(c) 3 (d)-5

21. The pay off I „ . I is strictly determinable, then the value of A. is

- 4 A, 1

(a) zero (b) positive value

(d) negative value (d) any value

r_5 21 player always uses

22. For the pay off ^ . the row

(a) first strategy

(b) second strategy
(c) mixed strategies of both first and second
(d) no strategy

23. In the above problem what will be better strategy for a colujin player

(a) first strategy

(b) second strategy

(c) mixed strategy with —, —

(d) mixed strategy with —, —

24. In the problem given in Q.No. 22, the value of the game is

(a) - 5 (b) 2
(c)-7 <d)-3

25. In the pay off of problem given in Q.No.22, the winner is

(a) row player (b) column player

(c) neither of the two (d) none of the above
614 Operations Research : Theory and Practice

26. Srija and Sanja play a game with two types of coins 5 ps and 10 ps. Each draw
one coin randomly and if the sum is even Srija wins the coins, otherwise Sanja.
The value of the game is

(a) 1/2 (b)5

(c) zero (d) — 5

27. In the above problem, the strategy used by Srija is

(a) 5 ps always

(b) 10 ps always

(c) both strategies with probability —, —

(\ 3
(d) mixed strategies of — —
28. In a game between cricket teams India and Pakistan the toss is supposed to play
major role. Both captains take one coin and toss. If both get same, India wins
otherwise Pakistan. If we take T for winning and -1 for losing, the value of the
game is

(a) 1 (b) -1
(c) zero

29. Identify the dominance in

(a) R^ dominates i? 2 (b) R2 dominates #3

(c) C3 dominates C2 (d) C2 dominates C3

30. Which of the following game is not fair

«[S 2] «»l 1
-1
-1
1

(C)
+ 10
+5
-10
1
(d) 0 °11
Game Theory 615

Fill in the Blanks

1. A game is said to be if the value of the game is zero.

2. If a game uses pure strategies, it is called

3. A graphical solution can be used for games if the order of the pay off is
or

4. If gain of a player is loss of the other then the game is called

5. Deterministic games use strategies

6. strategies are used in the games without saddle points

7. If every element in R± is less than or equal to corresponding element in R2,
then is said to be dominating.
8. According to basic assumptions of game theory, row player always wishes to
and column player wishes to

9. The value of the game whose pay off is a 2 x 2 identity matrix is

10. When a game is represented by LPP, if row player represents primal, column
player represents

11. A game with more than two players is called player game

12. If loss of a player is not a gain to the other, then the game is called
game
13. The game whose pay off is null matrix is . game

14. The best strategy for row player in the game with pay off is

15. The value of the game with pay of is

16. In the pay off 9 the saddle point is at and is equal

to
17. If minimax value is same as maximin value the game said to have

18. The corresponding strategy of each player at equilibrium point is

strategy
616 Operations Research : Theory and Practice

19. T h e value of the game whose pay off is L. .is

20. T h e list of all possible actions that will take for every pay off that might result
is called

Answers
Objective Type Questions :
l.(c) 2. (a) 3. (b) 4.(d) 5.(c)
6. (b) 7. (a) 8. (a) 9.(c) 10. (d)
11. (b) 12. (c) 13. (d) 14. (b) 15. (c)
16. (c) 17. (c) 18. (b) 19. (a) 20. (b)
21- (d) 22. (a) 23. (a) 24. (a) 25. (b)
26. (c) 27. (c) 28. (c) 29. (c) 30. (d)
Fill in the Blanks :
l.fair 2. deterministic
3. 2 x n or n x 2 4. zero sum game
5. pure 6. mixed
8. Maximise gains, minimise losses
10. dual

11. multi 12. non zero sum

13. fair 14. mixed strategy of both i ows with
(1/2, 1/2)
15.zero 16.(i2i,Ci), 5
17. saddle point 18. optimal
20. strategy
618 Operations Research : Theory and Practice

CHAPTER AT A GLANCE

DYNAMIC PROGRAMMING

Identify decision variables

Specify objective function

Decompose into smaller problems (stages)

Identify state variables

State the general recursive relations. Also

decide forward/backward

CaLaulate required values of return function

by appropriate formulae at each stage

Repeat & determine the overall optimal policy

Dynamic Programming 619

12.0 Introduction
Dynamic programming is an optimization technique of multistage decision
process. The word "dynamic" means to the situations of change in several stages, such
as every week, every day or every month etc., and 'programming' is the term used in
the mathematical sense of selecting an optimum allocation of resource.
Rechard Bellman propounded this technique of finding optimal solution by
developing recursive relation. In dynamic programming, a large problem is split into
smaller sub problems each of which involve in a few variables. These sub problems
are sequentially optimized so as to make the total problem optimal.
An important point to note here is that - the problem of successive stages is
treated separately even though by the very nature of the problem, these stages are
dependent. The answer to this question is based on "Bellman's principle of
optimality", which is stated as "An optimal policy (set ofdecisions) has the property that what
ever the initial state and decisions are, the remaining decisions must constitute an optimal policy
with regard to the state resulting from the first decision".
Discrete and continuous, deterministic as well as probablistic models can be
solved by this method. Thus dynamic programming method is very useful in obtaining
optimal solutions of various problems such as inventory, replacement, allocation,
linear programming, reliability calculations etc. A single constraint problem is
relatively simple, but in the problem of more than two constraints more complexities
may appear.

12.1 The Approach & Algorithm

The three basic features that are present in the structure of the Dynamic
Probramming Problem (DPP) problem are as follows.
1. The DP problem can be decomposed into sequence of smaller sub-problems
called stages of the original problem. At every stage there will be number of
decision alternatives and a decision is made by most suitable one out of given
list. Stages usually represent different time periods associated with planning
period of problem, places, people, or any other suitable parameter

e.g. In replacement problem, each year is a stage, in the sales man allocation problem,
each territory may be a stage.

2. The states are represented in various conditions of decision process at a stage.

The variables that specify the condition of decision process at a particular stage
are called states. These states can provide information for analysing the possible
effect that the current decision could have upon future action.

e.g. A specific city is referred to as state variable in any stage ofthe shortest root problem.

3. At each stage, a decision is made which can influence the state of the system at
the next stage to arrive the optimal solution at the current stage. This decision
described in the form of an algebraic equation to explain its worth or benefit,
is referred to as Return function.
620 Operations Research : Theory and Practice

The solution of a DPP is based on Bellman's principle of optimality (recursive

optimization technique) which states:
An optimal policy has the property that, whatever the initial state and initial decision,
the remaining decisions must constitute an optimal policy with regard to the state resulting from
first decision.
With reference to the above principle, the solution procedure goes by solving
sequentially adding a series of sub problems until the overall optimum is obtained. If
the problem is sequenced from last stage to first (i.e., initial stage), the method is said to be
backward recursion while U is said to be forward recursion if it is sequenced from first (i.e.,
initial) to the last.
General Algorithm :
Step 1 Identify the decision variables of the problem and specify objective function to
be optimised under the constraints, if any.
Step 2 Decompose the given problem into number of smaller problems (or stages).
Identify the state variables at each stage and write down the transformation
function as a function of state variable and decision variables at the next stage.
Step 3 State the general recursive relationship for computing the optimal policy.
Decide whether to apply the forward or backward recursion to solve the
problem.
Step 4 Construct appropriate tables to show the required values of the return function
at each stage as shown below.
Possible decision (d.a) F,.(s,,>dn) Optimal Return Optimal
entering states (SJ d,, K (s») Decision (d*n)

Step 5 ; Repeat and Determine the overall optimal policy or decision and its value at
each stage. There may be more than one such optimal policy also.

12.2 Salient Features & Their Definitions of DPP

On summarising the above topics we get the following salient features in DPP.
1. Recursive Function : A dynamic programming problem is solved by an
objective function which is recurring in nature. This recursive approach can be
applied in two directions.
(a) Forward Recursive Approach: In this approach the main problem is atacked
by dividing it from intial stage to final stage.
(b) Back Ward Recursive Approach : In this approach the problem is dealt by
sub dividing the problem from last to first and then solving from final to
intial stages.
2. Stage : A large problem is decomposed into smaller and sequential
sub-problems. Each sub-problems is referred to as 'a stage' where a decision is
called for. Thus each stage can be considered to have a starting and ending,
the ending of one stage will be the beginning of its next stage.
Dynamic Programming 621
Suppose an engineering students has to acquire good marks in his degree after
studying for four years. Each year is considered to be a stage. When a student
completes his first year successsfully, he is said to have ended 1st stage (and started
his second stage) having 1/4^ of his large problem is completed. Thus he moves to
optimise his returns in second stage i.e.,2TC^ year with similar objective function, but
this includes the results of Ist year, and after completion of his second stage he acquires
2/4 of problem solved. Similarly, he goes to 3rd stage, finishing 3/4M that should
include the results of Ist and 2nd year. Finally, after completing his fourth year, he
completes his total problem that includes results of his previous three stages.

1 year 2 year 3 year 4 year

Stage 1 Stage 2 Stage 3 Stage 4

FIGURE I I I : STAGES OF A DPP

3. State : Any stage just before starting or just after completion will have certain
status called ' the state' and the variable that links up two stages is called state
variable.

4. Return Function : The decision of a stage described in the form of an algebraic

equation with state variables to explain the worth or benefit (or cost is known
as Return function or transformation equation.

Stage 1
Stage 1 [Stage 2 Stage 2 Stage n

ith oth
Status:
Status: ~ Status: ~ Total
of problem of problem problem
is completed i 5 completed

FIGURE 1 3 . 2 : STAGES AND STATES IN A DPP

12.3 Characteristics of DPP

The basic the characteristics of the DPP are given below.
1. The problem can be sub-divided into sub-problems referred to as stages with
policy decision at each stage.

2. A stage is a device to sequence the decisions. This means that the sequential
summation of these decisions obtained from optimal solution of sub-problems
will be the part of the optimal solution of the total problem.
622 Operations Research : Theory and Practice

3. Every stage consists a number of states associated with it. The states are trie
different possible conditions in which system may find itself at that stage of the
problem.
4. Decision at each stage converts the current stage into state associated with the
next stage.
5. The state of the system at a stage is described by a set of variables called state
variables.

6. When the current state is known, an optimal policy for the remaining stages is
independent of the policy of the previous stages.
7. To identify the optimum policy for each state of the system, a recursive equation
is formulated with n stages remaining, given the optimal policy for each state
with (n - 1) stages left.

8. Using recursive equation approach each time the solution moves either
backwards or forwards by stage obtaining the optimum policy of each state for
that particular stage, till it attains the optimum policy beginning at final or
initial stage respectively.
fUUSTRATiOftl 1

A medical representative located at city has to travel to city 10. He knows the
distance of alternative routesfrom city 1 to city 10 and has drawn the network map
as shown in thefollowing figures along with the distance between the cities. Find
shortest possible route. Also find the shortest routes from any city to city 10.
[JNTU (Mech.) 96]

Stage 5. Stage 4 . Stage 3 . Stage 2 . Stage 1

FIGURE 123:

The above problem can be divided into five stages as shown with dotted lines
in the figure, and backward recursion is chosen with the objective to find shortest
route. It is solved in the tabular form method as given below.
At stage 1, if the medical representative is at city 10, it is the destination itself,
he need not travel any more.
Dynamic Programming 623

Stage 2 : At stage 2, i.e., if he is at city 8 or city 9, the shortest path calculation is shown
below.

^ \ ^ Decisions Destinations i.e., city 10 Minimum Distance Optimal Decision

(Optimal)
State^^ Stage Distance Cummulative
Variable^) ^ \ Distance
Entering 8 7 7 7 8 to 10
State g
9 9 9 9 to 10

Stage 3 : Now, using these optimal results of stage 2, we proceed to stage 3.

^ ^ Decisions Destinations Min. Optimal Cummul.

(Optimal) Stage Optimal
City 8 (7) City 9 (9) Distance Decision Decision
State ^ v
Variable (52) \ . Stage Cummul. Stage Cummul.
Distance Distance Distance Distance
Entering 5 4 11 8 17 11 5-8 5-8-10
Variables
6 3 10 7 16 10 6-8 6-8-10
7 8 15 4 13 13 7-9 7-9-10

Stage 4 : With the above optimal decision we compute stage-4 decisions.

(Figures in parantheses are the optimal distances derived from previous table).

\ Decisions Destinations Min. Optimal Cummu.

(Optimal) Stage Optimal
City 5 (11) City 6 (10) City 7 (13) Distance Decision Decision
State \ . Stage Cummu. Stage Cummu. Stage Cummu.
Variable (S3) \ Dist. Dist. Dist. Dist. Dist. Dist.
Entering 2 7 18 10 20 5 18 18 2-5 2-5-8-10
States or or
2-7 2-7-9-10
3 3 14 8 18 4 17 14 3-5 3-5-8-10
4 6 17 10 20 5 18 17 4-5 4-5-8-10

Stage 5: With the optimal (18,14 and 17) distance decisions we now proceed to calculate
final decision table of stage 5..

\v Decisions Destinations Min. Optimal Cummu.

(Optimal) Stage Optimal
City 2 (18) City 3 (14) City 4 (17) Distance Decision Decision
State\v Stage Cummu. Stage Cummu. Stage Cummu.
Variable (S4K Dist. Dist. Dist. Dist. Dist. Dist.
1 4 22 6 20 3 20 20 1-3 1-3-5-8-10
or or
1-4 l_4_5_8-10
624 Operations Research : Theory and Practice

Thus two optimal solutions are obtained as follows.

« , . , [sequence (city)(T) -» C$) -> f5) -» Cs) 10

S0luU0nSetl = 20
{Distance^) ~ 6 ~ 3 ^ 4 "

Solution set 2 {sequence (city)® ^ G > -*(t> - (D 10

= 20
1 Distance (km) 3 6 4

We can thus state that medical representative may have to travel, min. of 20
km from city 1 to city 10. If he is at other stations, the following optimal decisions may
be made

Stations Decision Min. Distance to travel Decision taken at

1 1-3-5-8-10 20 km Stage - 5
or
1-4-5-8-10
2 2-5-8-10 18 to Stage - 4
or
2-7-9-10
3 3-5-8-10 IA km Stage - 4
4 4-5-8-10 17 km Stage - 4
5 5-8-10 Ukm Stage - 3
6 6-8-10 10 km Stage - 3
7 7-9-10 13 km Stage - 3
8 8-10 7 km Stage - 2
9 9-10 9 km Stage - 2

ILLUSTRATION 2
Suppose the above problem is to maximise the distance assuming that he gets
more allowance if travels more. Now the decision tables will be asfollows.
Stage -2:

^^\^ Decisions Destinations city 10 Maximum Distance Optimal Decision

(Optimal)
State^^ Stage Distance Cummulative
Variable (Sx) N ^ Distance
Entering 8 7 7 7 8- 10
State
Variables 9 9 9 9 9 - 10
 Dynamic Programming 625
Stage 3 :

^ ^ Decisions Destinations Max Overall

(Optimal) Optimal
City 8 (7) City 9 (9) Distance Decision
State ^ v Stage Cummul. Stage Cummul.
Variable (S2) \ . Distance Distance Distance Distance
Entering 5 4 11 8 17 17 5-9
Variables
6 3 10 7 16 16 6-9
7 8 15 4 13 15 7-8

Stage 4 :

\^ Decisions Destinations Max. Optimal Overall

(Optimal) Stage Decision
City 5 (17) City 6 (16) City 7 (15) Distance Decision
State \ ^ Stage Cummu. Stage Cummu. Stage Cummu.
Variable ( S 3 ) \ Dist. Dist. Dist. Dist. Dist. Dist.
Entering 2 7 24 10 26 5 20 26 2-6 2-6-9-1J0
States
3 3 20 8 24' 4 19 24 3-6 3-6-9-10
4 6 23 10 26 5 20 26 4-6 4-6-9-10

Stage 5:

\. Decisions Destinations Max. Optimal Overall

(Optimal) Stage Decision
City 2 (22) City 3 (24) City 4 (26) Distance Decision
State \ v Stage Cummu. Stage Cummu. Stage Cummu.
Variable ( S 4 ) \ Dist. Dist. Dist. Dist Dist. Dist.
1 4 30 6 30 3 29 30 1-2 1-2-6-9-10
or or
1-3 1-3-6-9-10

Here also we get two solutions as given below :

„ , . , [sequence (city) (1 ) -» ( 2 ' -> ( 6) -» (9)
(9) -> (1Q> „„
SolutlOn set 1
{Distance (L), 4 10 ^ 7 9 = 3°
Solution set 2 = 30
' 1 Distance (km) 6 8 7 9
(Other stage decisions can be read at the respective tables).
Total max. distance, the medical representative can travel upto 30 Aw without
repitition.
626 Operations Research : Theory and Practice

ILLUSTRAION 3
Find number of each of three items to be included in a package so that value
of package will be maximum. Total weight of package must not exceed 5 Kgs.

Item Weight in Kgs Value in Rs.

1 1 30
2 3 80
3 2 65

[JNTU (CSE) 98]

Solution
Hie given problem can be formulated as follows:
Let xlt x2 and x3 be the number of items of each type
maximise f(x)= 30 x± + 8(k2 + 65 x% (The total value is to be maximised)
subject to xx + 3x2 + 2x3 < 5
xx, x2 , x3 > 0
We have to determine how many units of three types of items are to be loaded.
So it is a three stage dynamic programming problem (DPP). Let:*. (/ = 1 , 2 , 3) denote
three decisions. Let £ (x ) denote the value of the optimal allocation for the three types
of items.
If fj(S ,Xj) be the value associated with the optimum solution
£ (s), (j = 1, 2, . . . n) then we have
/ (5) = m a x (5 , Xj)

and j (s) = m a x {py (Xj) +/*_ Y (s - Xj)}

0<<
0<x{<s

j= 1,2,3
where pj (x ) denotes the expected value obtained from allocation of #• units of
weight to the item/
Now, for stage -1 problem (i.e., for one item loading).

= max{30 xx)

where the largest value of Xj = — = 5

(Here W is. total weight and w^ is weight of each of item 1.)

Hence we have 6 alternatives i.e., 0, 1, 2, 3 , 4 a n d 5 for x^
We have the following computations.
Dynamic Programming 627

Value of 30 xx Optimum solution

s\ 0 1 2 3 4 5 fi(S) x\

0 0 0 0
1 0 30 30 1
2 0 30 60 60 2
3 0 30 60 90 90 3
4 0 30 60 90 120 120 4
5 0 30 60 90 120 150 150 5
For 2-stage problem, the largest value of x% is ["5/3"] = 1 and thus we have two
alternatives i.e., 0 and 1 for x2

f2 (s) = max {80 x2 +/J (S - 3 x2 )}.

The computations in the 2-stage problem are as follows.

Valueof80x 2 +/^(5-3x 2 ) Optimum solution

s\ 0 1

0 0+0=0 0 0
1 0 + 30 = 30 30 0
2 0 + 60 = 60 60 0
3 0 + 90 = 90 80 + 0 = 80 90 0
4 0 + 1 2 0 = 120 80 + 30 = 110 120 0
5 0 + 150 = 150 80 + 60 = 140 150 0
For 3-stage problem the largest value of x$ is [5/2] = 2 and so we have 0, 1 and
2 as the alternatives for x3

And, / 3 (5) = max {65 xs +f2 (S - 2x$ )}

The computations of 3-stage problem are as follows.

V X2 Value of 65 x$ +f$ (S - 3x3 ) Optimum solution

s\ 0 1 2
4
0 0+0=0 0 0
1 0 + 30 = 30 30 0
2 0 + 60 = 60 65 + 0 = 65 65 1
3 0 + 90 = 90 65 + 30 = 95 95 1
4 0 + 120 = 120 65 + 60 = 125 130 + 0 = 130 130 2
5 0 + 150 = 150 65 + 90 = 155 130+30=160 160 2
628 Operations Research : Theory and Practice

For the given total weight W = 5, the optimum solution is ,

x\ = 1 , x\ = 0 , x% = 2

andJl (s) = 160 is the maximum value of load. In other way,

Item No. Qty. to be Loaded Unit Value (Rs.) Total Value (Rs.)
1 1 30 30
2 0 80 0
3 2 65 130
Grand total 3 Rupees 160

ILLUSTRATION 4
Solve the following using tabular method of solution. The owner of a chain of four
grocery stores has purchased six crates of fresh fruits. The estimated probability
distributions of potential sales of the fresh fruits before spoilage differ among the
four stores. The following table gives the estimated total expected fruits at each
store, when it is allocated to various number ofcrates:
Number of Stores
Crates
1 2 3 4
O O O O O
1 4 2 6 2
2 6 4 8 3
3 7 6 8 4
4 7 8 8 4
5 7 9 8 4
6 7 10 8 4

For administrative reasons, the owner does not wish to split crates between stores.
However, he is willing to distribute zero crates to any of his stores. Find the
allocation of six crates tofour stores so as to maximise the expectedprofit
[JNTU(EEE)98/OT]
Solution \
Let the four stores be considered as four stages in a dynamic programming
formulation. The decision variables xy (/ = 1 , 2 , 3 , 4) denote the number of crates
allocated as the f1 stage from the previous one.
Now, letft(x:) be the expected profit from allocation of X: crates to store/ Then
the problem can be formulated as an L.P.P as follows:
Maximize Z = p\ (x±) +/>2 (X2 ) +Ps (X3 ) +PA (X4 )
Subject to the constraints
x +X +X +X =
l 2 3 4 6>
Dynamic Programming 629
Let there be S crates available for j remaining stores and x, be the initial
allocation. Define/- (x-) as the value of the optimal allocation for stores 1 to 4 both
inclusive.
Thus for stage j = 1, fx (S , xx) = {pj (xx)}
If f:(S,Xj) be the profit associated with the optimum solution
Jj (S) (j = 1 , 2 , 3 , 4) then
f{(S) =Max px(xx)
0<Xj<5

Thus the recursive relation is

for j = 1, 2, 3,4
and fj (S) = max fa {xj ) +JJ-+ 1 (S-Xj )}
0 < Xj;< S.

The solution to this problem starts with 74 (S) and is completed when/J (5) is
obtained. (Backward recursion)
The computations for stage -1 problem (i.e., for j = 4) are as follows :
s A *4
0 0 0
1 2 1
2 3 2
3 4 4
4 4 3,4
5 4 3,4,5
6 4 3,4,5,6
Forj = 3, we have a two-stage problem. The computations are shown below :

/s(S,*s> Optimal Solution

s \ 0 1 2 3 4 5 6

0 0+0 0 0
1 0+2 6+0 6 1
2 0+3 6+2 8+0 8 1,2
3 0+4 6+3 8+2 8+0 10 2
4 0+4 6+4 8+3 8+2 8+0 11 2
5 0+4 6+4 8+4 8+3 8+2 8+0 12 2
6 0+4 6+4 8+4 8+4 8+3 8+2 8+0 12 2,3
630 Operations Research : Theory and Practice

For j - % we have three stage problem. So we have

=
/2(S,*2)
0 1 2 3 4 5 6
s \ ^
0 0+0 0 0
1 0+6 2+ 0 6 I
2 0+ 8 2+ 6 4+0 8 0, 1
3 0+10 2+ 8 4+ 6 6+0 10 0, 1,2
4 0+11 2 +10 4+8 6+ 6 8+0 12 1,2,3
5 0+12 2 +11 4+10 6+ 8 8+6 9+0 14 2,3,4
6 0+12 2 +12 4+11 6+10 8+ 8 9+ 6 10+0 16 3,4
For j = 4, we have the required four-stage problem :

/i(S.*i) Optimal Solution

S N^ 0 1 2 3 4 5 6
*i
6 0+16 4+14 6+12 7+10 7+8 7+6 7+0 18 1,2

From above computations, it is observed that the maximum profit of Rs. 18 can
be obtained by choosing the following eight alternative solutions :

Store 1 orxj Store 2 or x\ Store 3 or #3 Store 4 or x\

1 2 2 1
1 3 1 1
1 3 2 0
1 4 1 0
2 1 2 1
2 2 1 1
2 2 2 0
2 3 1 0

ILLUSTRATION 5
Solve thefollowing LX.P by dynamic programming method:
Maximize Z=2xI + 5x2
subject to 2xj +x2< 430,2x2 < 460

[JNTU (CSE) 97/S]

Dynamic Programming 631

S o I uf ion :
Max Z = 2xj + 5 x2 subject to 2 xj + x2 ^ 430.
2x2<460;x1 , x2>0
The L.P problem can be considered as a two-stage, two-state problem because
there are two decision variables* and two constraints*. Starting with second stage
backward recursion, the procedure is as follows:
The optimal value off2 (c± , c2 ) at the second stage is given by
f2(c{,c2)=Max {5x2}
0<x2<c
where q = 430. c2 = 460. The feasible value of x2 is a non-negative value which
satisfies all the given constraints x 2 < c 1 ( = 430) and 2x 2 < c2 (=460).Thus the
maximum value V that x2 can assume is; c - Min 1430, ~^~> = 230.
Therefore^ (cj , c2 ) = Max {5 x2}
0 < x2 < c

= 5 min {430 -2xx ; 230}

and x\ = Min {430 - 2 xh 230} = 230
Proceeding backwards to stage 1 (j = 1), the recursive relation for optimizaiton
can be expressed as :
/i (q , c2 )=Max {2 *! + ^ (Cl - 2xx), 230)}

= Max [2*! + 5 min {430-2x^230}]

Where maximization variable xj, satisfying the conditions

2 xj < c± (= 430) and oxj < c2 (= 460) is minimum of
_.. [430 460]
c = A f m j — , — I = 215.
Since the minimum (i.e., 230) of (430-2*!, 230) is obtained at x{ = 100 for
< 215 , we get
fx (Cl , c2 ) = Max [2*! + 5 Afm {430-2x i; 230}]
= M a x [ 2 x 100 + 5 x 2 3 0 ]
= 1350
and x* =100
Hence the optimum solution to the given L.P. problem is Xj = 100; x2 = 230
and Max. Z^ 1350.
* The number of decision variables of LPP represent the stages of DPP and the number
of constraints of LPP correspond to the states of DPP
632 Operations Research : Theory and Practice

ILLUSTRATION 6
A firm has accepted to supply 75 bulbs at the end of the first month and 115 bulbs
at the end ofthe second month. The cost ofmanufacturing any bulb in any month
is given by 45x + 0.2x rupees where x is the number ofbulbs manufactured in the
month. However the company can manufacture more number of bulbs and
carryout to the next month, and, an inventory carrying charges ofRs. 8/- per bulb
is to be charged. Find the number of bulbs to be manufactured in each month to
maintain the total cost to minimum. (Use calculus method ofsolution), (no initial
inventory).
Solution:
Let us first formulate the NLP problem.
Let xi = No. of bulbs to be manufactured in first month
x2 = No. of bulbs to be manufactured in second month
Objective Function:
Now, manufacturing cost in first month = 45xj + §J2x\
and manufacturing cost in second month = 45x2 + 0.2v|
and there is no initial inventory, hence no carrying cost in first month.
But, the carrying cost in second month, if x1 exceeds 75,
= 8(*!-75)
Hence total cost = 45xj + 0.2*f + 45x2 + 0.2*f + 8 (x\ - ?5)
This total cost is to be minimised. Therefore objective function is
Minimise Z = 0.2 x\ + 0.2 x\ + 53 xx + 45 x2 - 600
Constraint Set: According to the agreement of supply
xj > 75 [Requirement constraint on agreed supply for first month]
xx + x2 = 190 [Requirement constraint on agreed supply for first two months]
Conditions of Variables: Since the production of negative number of bulbs is
meaningless, we have non-negative conditions for both variable and also can not be
fractions
xY > 0, x2 > 0
Summary: The problem is summarised below
Min Z = (45*! + 0.2*f) + [(45x2 + 0.2x|) + 8 (xi - 75)]
subject to xx > 75
xx +x 2 = 190 and xj > 0, x% > 0 and integers
Solution by Calculus Method:
To solve the above problem, we start from second month and go by backward
recursion. If/2 is the inventory at the beginning of the second month, the optimum
number of bulbs to be manufactured in the second month is given by
4 = 115-/ 2 . . . . (i)
and the cost incurred in the second month by
r2{x%I2) = 8/2 + 4 5 4 + 0 . 2 4 2 . . . . (ii)
Dynamic Programming 633

By using equation (i), R% can be expressed as

R2 (/2) = 8 /2 + 45 (115 - 72) + 0.2 (115 - /2)2
= 8 /2 + 5175 - 45 /2 + 0.2 (13225 +1\ - 230 /2)
= 0.2 /| - 83 72 + 7820
Since the inventory at the beginning of the first month is zero, the cost involved
in the first month is given by
RI(XJ) = 45 XJ +0.2*|
Thus the total cost involved is given by
/2(/2, *!)= (45 *i + 0.2 x?) + ( 0 . 2 / f - 8 3 / 2 + 7820) . . . . (iv)
But the inventory at the beginning of second month is related to x j as
78 - *! - 75 . . . . (v)
Equations (iv) and (v) lead to
/=/2<'2>
= (45 *j + 0.2 xf) + |~0.2'(*i - 75)2 - 83 (*i - 75) + 7820l
= 0.2 x\ + 0.2 (xf - 150 ;ct + 5625) + 45 K J - 8 3 x 1 + 6225+7820
i.e., /=/(*!) = 0.4 xf-68x1+ 15170
Since/is a function of x j only, the optimum value of x^ can be obtained by
calculus methods as
-I- =— [0.4 «?- 68 *, + 15170] = 0
<&! dx^ L J
CO

= 0.8x 1 -68 = 0 or x, = - — = 8 5
O.o

As ^4 = °-8 > °>

<£x{
the value of «j corresponds to the minimum value of/
Thus the optimum solution is given by
105
x* = 85; *2 =
and
,/min =/(3i;i) = °'4 (85)2 - 68 (85) +15170
= Rs. 12.280/-
[Manufacturing cost is Rs. 5270/- in the first month and Rs. 6390 in the second
month while the carrying cost on excess (10 bulbs) manufactured in the first month
is Rs. 80/-. Total Rs. 12.280/-]
ILLUSTRATION 7
Use Bellman's principle of optimality to find the optimum solution to the
following problem:
Minimize Z=y2, +y22 +y33 subject to the constraints
y,+y2+y,>is y,,y»y3>o
[JNTU (EEE)
MNTIJ fEEEl 98/OT, (CSE) 2000
98/OT. rCSE"! 2000 JNTU
.INTII Prod/Mech/Chem/Mechatronics 2001/S]
Prnd/Medi/rhcm/Mcrliatrnnirs - 2001/SI
634 Operations Research : Theory and Practice

i?olwfio>-\ !
Since the decision variables are y \ , y % andy3 the given problem is a 3 stage
problem defined as follows :
53 = J i + 3 > 2 + j ) > 3 > 1 5
S 2 =3>i+3> 2 = S3-3-3 and
s
l = y\ = s2-yz
Therefore the functional (recurrence) relation is
/1(51) = m i n ^ = ( s 2 _ ^ ) 2
OS^SS,

/2 (S2) = min J3>f + 3!} = min \y\ + /i (S2)}

0<3i 2 <S, 0<j> 2 <S 2
and /3(S3) = min \y\ +y\ + ji§| = min {j>l+/ 2 (S 2 )j
0<j 3 <S 3 0<ji 3 <5 3
2 2
/2 (S2) = mm [yl + (S 2 -j 2 )2J - ft S21 + L - | S21
OSJPJSSJ r J I ^ J
1
~- 2 SS22
According to maxima -minima principles, the minimum value of function
y\ + (S^-y^ can be obtained by equating the first derivative to zero
^f^-^j-o
i.e. 2y2-2(S2-y2) = Q

=> 4>2 = 2 S 2 •'• yz=2S2

/3 (Sj) = min (,§ + /2 ( S2 )| = min Lf + ^ (S3 -3-3 )2}

0<^<5 3 0< > 3 <S S I ' i
or /3 (15) = mm fy| + \ (15 -j»8 )2}
0<>- 3 <S 3 I " J
since, Ss(yi+y2+y^) ^ 15
Since me minimum value of the function 313 + — (15 -y$ )2 occurs at 313 = 5 , we have

/3(15)=J5 2 + |(15-5)2U75
Thus, S3 = 15 implies that 313 = 5;
S2 = Sg -3)3 = 1 5 - 5 = 1 0 implies that
3>2 = g S2 = 5
Sj = S2 - j2 = 1 0 - 5 - 5 implies that
y\ = Sn = 5
Hence the optimal policy 133^ = 5, y% = 5 and313 = 5 with/3 (^) = ^-
Dynamic Programming 635

ILLUSTRATiOfli 8

Using dynamic programming approach, solve reliability problem with following

data. | 1 1 1 1
i=l i=2 i=3
Mi —i— —.— —i—
R C R C R C
1 0.6 2 0.8 3 0.7 1
2 0.7 4 0.8 5 0.8 2
3 0.9 5 0.9 6 0.9 3
The total capital available is 10 (in units of thousand rupees)
[JNTU (Mech.) CCC-2001/S, (ECE) CCC-2001/S]

Solufion :

Let R be the total reliability of a system of 'ri components arranged in series

and let there be 'k' parallel units per component (j - 1,2, 3,)

Now, we formulate the problem as

Minimise R-R^ ; R% ; R3

Subject to constraints Cj + C2 + C3 < C where 'C' is total capital available i.e., 10

Now, to develop the recursive relationship.

Let us define the following.
X:: The capital allocated to the stage j where j= 1, 2, 3.
kj : Number of parallel units assigned onto main component (/')
f: (x:) : Reliability of components (stages) 1 toj given that 0 < x< C

The recursive equations are :

/ 1 (x 1 ) = max {^(C!)] ; * i = l , 2, 3
0<Cl<xj

and fj (Xj) = max fa (C-) ; f j ^ l (Xj- C;-)}; j = 1, 2, 3 and kj = 1, 2, 3

0<Cj<Xj
Starting with stage-1, since main component-1 must include atleast one
(parallel) unit, we find that x} must at least equal C: (1) = 2. By the same reasoning
Xj cannot exceed 10 - (3 + 1) = 6; otherwise remaining capital will not be sufficient to
provide main components 2 and 3 with atleast one (parallel) unit each. Following the
same reasoning we get x% = 5, 6, 7, 8 or 9 and xg = 6, 7, 8, 9 or 10.
636 Operations Research : Theory and Practice

Stage 1: /j (xj) = max. R^ (C,)

Xj Aj = 1 &j = 2 Aj = 3 Optimal Solution

R, = 0.6 ; Cl = 2 /?! = 0.7, C, = 4 \R, = 0.9 ; C, = 5 /} (KH) ^
2 0.6 0.6 1
3 0.6 0.6 1
4 0.6 0.7 0.7 2
5 0.6 0.7 0.9 0.9 3
6 0.6 0.7 Q.9 0.9 3

Stage 2 : fo (x2) = R% (C2)/i (x2 - C2)

*2 kl = 1 A! = 2 /^ = 3 Optimal Solution
/?2 = 0 . 8 ; C 2 = 3 7?2=0.8, C 2 = 5 J?2 = 0 . 9 ; C 2 = 6 /2 (*2) £*

5 0.8x0.6 = 0.48 0.48 1 !

6 0.8x0.6 = 0.48 0.48 1
7 0.8x0.7 = 0.56 0.8x0.6 = 0.48 0.56 1
8 0.8x0.9 = 0.72 0.8x0.6=0.48 0.9x0.6 = 0.54 0.72 1
9 0.8x0.9=0.72 0.8x0.7 = 0.56 0.9x0.6 = 0.54 0.72 1

Stage 3 : fs (*3) = fl3 (C3)/2 (X3 - C3)

xs k^ = 1 kl = 2 kl = 3 Optimal Solution
/?3 = 0 . 7 ; C 3 = 1 *3=0.8,C3 = 2 ^ = 0.9^=3 /g (*3) A*
6 0.7x0.48 = 0.336 0.336 1
7 0.7x0.48 = 0.336 0.8x0.48 = 0.384 0.384 2
8 0.7x0.56 = 0.392 0.8x0.48 = 0.384 0.9x0.48 = 0.432 0.432 3
9 0.7x0.72 = 0.504 0.8x0.56 = 0.448 0.9x0.48 = 0.432 0.504 1
10 0.7x0.72 = 0.504 0.8x0.72 = 0.576 0.9x0.56 = 0.504 0.576 2

Practice Problems
Solve the following LPP by Dynamic programming.

1. Max Z = &JCj + 7 *2. Subject to 2#i + *2 s; 8, 5* j + 2x2 < 5, x1; x2 ^ 0

Answer: xj = 0, *2 = 7.5 Max. Z = 52.5
Dynamic Programming 637
2. Max Z = 3*] + 7x2 , Subject to x\ + 4x% < 8, x2 < 2, xj , *2 > 0
Answer : xj = 8, x2 = 0, Zmax = 24

3. Max Z = 50 x{ + 100 x2 , Subject to 2xl + 3*2 < 48 , xl + 3x2 < 42 ,

x\+x%& 21, *i, *2 ^ 0

Answer : xj = 6, x:2 = 12, Zmax = 60

4. Min Z = *i + 3x9 + 4x3 , Subject to 2*j + 4^2 + 3x3 > 60 , 3xj + x2 + 3*3 > 90
x 1 ,x 2 ,x s ^0

5. Max. Z = 2*i + 5x2 . Subject to 2^ + x2 < 43 , 2x2 < 46, xj, x 2 > 0

Answer : x^ - 10, x2 = 23, Z max = 135

6. Suggest the best combination of advertising mdeia and advertising frequencies

to a shaving blade manufacturing company interested in selecting the
advertising, mdeia for product with help of following estimated data.

Frequency / Expected Sales in (Rs.) Television

Month Radio Newspaper
1 150 125 220
2 250 200 275
3 300 225 370
4 310 250 340

The cost of advertising in newspaper, radio and television are Rs. 450, Rs.
1000, Rs. 1800 respectively for one time appreance. The budget allocated for
advertisting is Rs. 5000/- month.

7. Find shortest path from vertex^ to B along arcs joining the verious vertices
lying between A and B. The lengths of paths are as shown here.

FIGURE 12,4:
638 Operations Research : Theory and Practice

8. Find the shortest path from vertex A to B given. The lengths of paths are shown
in the following network.

FIGURE 1 2 . 5 :

9. Using dynamic programming find the shortest path from city 1 to city 10 in
the following network. The distance between two cities is given on arrow
connecting, them.

FIGURE 1 2 . 6 :

10. The routes of an air line, which connects 16 cities (A ,B, P) are shown in
fig. below. Journey from one city to another is possible only along the lines
(routes). Shown with the associated costs indicated on the path segments. If a
persons wants to travel from city A to city P with minimum cost, without any
back tracking, determine the optimal path (route) using dynamic
programming.

FIGURE 12.7 :
Dynamic Programming 639

11. A minimum cost pipe line is to be laid between points (towns) A and E. the pipe
line is required to pass through one node out of 5j, B% and Bs one out of
C], C2 and C3 and one out of Dj, D2 and D3 (see fig. below).

FIGURE 12.3 :

12. Find the minimum path by solving the dynamic programming problem given
below.

FIGURE 119:

13. Use dynamic programming to find the value of. Max Z^y^y^y^, Subject to
constraint : y i + y% + y^ = 5 and y^y? ,y$ > 0

Hints : f{ (X]) = max (z t ) and/-(ay) max {z/-J^_ l (x-- zj^,j = 1, 2, 3.

osz ; s* ; .

(5 5 5} (5^
Answer : -, - - with/3 (5) = -
I ' J ( )
14. Use dynamic programming to Max Z-2xl + 3x2 Subject to constraints

X j - x2 < 1, x j + jcg ^ 3, X[ + x2 > 0 and Xj, x2, x3 > 0

Answer : Xj = 0 , x2 = 3 , Max Z = 9

15. Use DPP method to Min Z = x2 + 3x2 + 4x3, Subject to 2xj + 4x9 + 3x3 > 60 ,
3x1 + 2x2 + x 3 > 6 0
23Cj + x2 + 3x3 > 90 and x j , xg , xs > 0
640 Operations Research : Theory and Practice

Review Questions
1. Explain the salient features of DPP. [JNTU (Mech.) 96/CC, ECE/CC - 96, (Mech) 94]
2. Write a short note on Dynamic Programming for Optimization.
[JNTU (CSE) 97, (Mech) 98/S, Mech 99]

3. Define Bellman's principle of optimality and its application to DPP.

[JNTU (EEE) 97/S, Mech/Prod/Chem/Mechatronics - 2001/S]

4. Write short notes on

(a) stage (b) state (c) recursive approach [JNTU (CSE) 96/S]
5. Explain tabular method of solving DPP [JNTU (EEE) 98, (Mech.) 95]
6. Write a short note on the characteristics of DPP [JNTU - Mech. 94/S]
7. Explain step by step procedure for developing an optimal decision policy in
Dynamic Programming Problem with an example. [JNTU (ECE) 97]

Objective Type Questions

1. In dynamic programming, which of the following does not change at every stage
(a) objective function (b) recursive function
(c) results (d) state variables
2. An optimal policy (set of decisions) has the property that whatever the initial
state and decision are, the remaining decisions must constitute an optimal
policy with regard to the resulting from the first decision
(a) state (b) function (c) recursion (d) stage
3. The sub problem of a main problem in a DPP is said to be
(a) state (b) function (c) recursion (d) stage
4. An LPP solved by using DPP method assumes the variables as
(a) states (b) stages
(c) recursions (d) return functions
5. Which of the following is considered as recursive optimisation
(a) linear programming problem (b) dynamic programming problem
(c) assignment problem (d) transportation problem
6. The optimality principle ( to solve DPP) is given by
(a) Dantzig (b) Richard Brown
(c) Bellman (d) Kimball
7. The conditions prevailing in decision process at a sub-problem of a main
problem are said to be
(a) stages (b) functions
(c) states (d) recursions
Dynamic Programming 641
8. A decisions that is made has its own benefits represented in an algebraic
equation is called
(a) stage (b) state variable
(c) return function (d) decision variable
9. Return function of DPP depends on
(a) state variables (b) stage decision
(c) decision policy (d) all the above
10. The equation that converts all the stages of the DPP is known as
(a) state function (b) stage function
(c) transition function (d) return function
11. The final stage decision is first solved in recursion
(a) forward (b) backward
(c) upward (d) downward
12. A DPP will have
(a) no objective function
(b) only one objective functions
(c) objective function equal to that of no. of stages
(d) infinite objective functions
13. A three variable, 2 constraint LPP if solved in dynamic programming approach
will have stages
(a) 2 (b)3 (c)6 (d)5
14. If an LPP has 'n' decision variables in '771' constraints, then it can be converted
into DPP with
(a) m stages, m states (b) m stages, n states
(c) n stages, m states (d) n stages, n states
15. Which of the following is not true in the case of DPP approach
(a) main problem is split into sub problems
(b) recursive relation ship is used
(c) values of return function remain same at all stages
(d) there may be more than one optimal policy
16. Find odd man out with reference to dynamic programming
(a) objective function (b) bijective function
(c) recursive function (d) transformation function
17. Characteristics LPP and DPP approaches differ in
(a) objective function (b) constraints
(c) variables (d) conditions of variables
18. DPP can not be applied to
(a) production scheduling problems (b) inventory problems
(c) net work analysis (d) behavioural problems
642 Operations Research : Theory and Practice

19. According to optimality principle, if an optimal solution is obtained to a system

of it must be optimal
(a) some portion (b) final stage
(c) any portion (d) at least one stage
20. When input and output are given as fixed which recursion works well ?
(a) forward (b) backward
(c) both recursions are equally good (d) no recursion can work well

Fill in the Blanks

1. An optimal policy has the property that whatever the and
are, the remaining decision must constitute and optimal policy with regard to
the stage resulting from the
2. A DPP is decomposed into smaller sub problems called
3. The conditions of a sub problem are known as
4. The algebraic equation that represents the benefits of a decision in DPP is called

5. The optimality principle is given by

6. The equation that interconnects all the stages of DPP is known as
7. An LPP of 'm' decision variables in 'n' constraints can be converted to DPP with
/ stages and stages
8. First stage decision is solved last in recursive approach
9. According optimality principle, if an optimal solution is obtained to a system,
it must be optimal to of the system.
10. A 3 - variable, 2 constraint LPP can be solved in DPP approach of
states and stages.

Answers
Objective Type Questions :
LJbj 2. (a) 3. (dl 4. (b) 5. (b)
6. (c) 7.(c) 8. (c) 9. (d) 10. (c)
11. (b) 12. (c) 13. (b) 14. (c) 15. (c)
16. (b) 17. (a) 18. (d) 19. (c) 20. (c)
Fill in the Blanks :
1 . Initial state, decisions, first decision 2. stages
3. states 4. return function
5. Richard Bellman 6. Transition function or state transformation
function
7. m, n 8. backward
9. any stage or portion 10. 2, 3