OPERTI

ONSRESEARCH
CHAPTERI
I
NTRODUCTION
1.
1.DEFI
NITI
ON
Oper
ati
ons Resear
ch i
s def
ined by many aut
hor
sin def
erentway
s.Howev
er,
f
oll
owi
ngdef
ini
ti
oni
spr
oposedher
e.

Oper
“ ati
ons Resear
ch i
s a sy
stemat
ic anal
ysi
s ofa pr
obl
em t
hrough sci
ent
if
ic
met
hods ,
car
ri
ed outby appr
opr
iat
e speci
ali
sts,wor
king t
oget
heras a t
eam
,
const
it
utedatt
hei
nst
anceofmanagementf
ort
hepur
poseoff
indi
nganopt
imum
andt
hemostappr
opr
iat
esol
uti
on,
tomeett
hegi
venobj
ect
iveunderagi
venset
const
rai
nts.
”

Meani
ngofOper
ati
onsResear
ch:
Fr
om t
heconceptanddef
ini
ti
ongi
venabov
e,Oper
ati
onsResear
chi
s:
1.Theappl
i
cat
ionofsci
ent
if
icmet
hods,t
echni
quesandt
ool
stot
hepr
obl
em t
ofi
nd
ananswer
2.Amanagementt
ool
int
hehandsofamanagert
otakeadeci
sion
3.Asci
ent
if
icappr
oacht
odeci
sionmaki
ngpr
ocess
4.An“
appl
iedr
esear
ch”ai
msatf
indi
ngasol
uti
onf
orani
mmedi
atepr
obl
em f
aci
nga
soci
ety
,indust
ryorabusi
nessent
erpr
ise.
Thi
si f
snot“undament
alr
esear
ch”
5.Adeci
sion-
ori
ent
edr
esear
ch,usi
ngsci
ent
if
icmet
hods,f
orpr
ovi
dingmanagement
aquant
it
ati
vebasi
sfort
aki
ngdeci
sionr
egar
dingoper
ati
onsunderi
tscont
rol
6.Appl
i
eddeci
siont
heor
y.I
tusessci
ent
if
ic,
mat
hemat
icalandl
ogi
calmeanst
otake
deci
sions.

1.
2.DECI
SIONMAKI
NG
Maki
ngappr
opr
iat
edeci
sioni
sthemostVI
TALaspecti
nmanagement.
Ever
yoneof
ust
akesanumberdeci
sionsev
eryday
.Somear
eimpor
tant
;somear
etr
ivi
al.Some
deci
si
onsi
nit
iat
easetofact
ivi
ti
es;someputanendt
o acer
tai
nact
ivi
ti
es.I
n
busi
nessenv
ironmentr
ightdeci
sionsatt
her
ightt
imesensur
esuccess.Thi
sshows
t
hei
mpor
tanceofdeci
sionmaki
ng.
Pr
obl
em i
s any v
ari
ati
on bet
ween what was pl
anned and what i
s act
ual
l
y
hav
e/pr
oduced.

-
Probl
em sol
vi
ngcanbedef
inedast
hepr
ocessofi
dent
if
yingadi
ff
erencebet
ween
someact
ualandsomedesi
redst
atesofaf
fai
rsandt
hent
aki
ngact
iont
oresol
vet
he

1
di
ff
erence.

-
Deci
sionmaki
ngr
equi
resf
oral
lhumanbei
ngbecauseeachofusmakedeci
sion
ev
erydayi
nourl
i
fe.
Thus,
deci
sionmaki
ngi
suni
ver
sal
.Deci
sionmaki
ngi
sar
ati
onal
sel
ect
ionamongal
ter
nat
ives.

Def
ini
ti
on
Deci
sionmaki
ngi
sthepr
ocessofsel
ect
ingorchoosi
ngbasedonsomecr
it
eri
a,t
he
bestal
ter
nat
iveamongal
ter
nat
ives.

Thedeci
sionmaki
ngpr
ocess:
St
epsi
nthepr
ocessofr
ati
onal
deci
sionmaki
ng
1.I
dent
if
yanddef
inet
hepr
obl
em
Pr
obl
em i
sanecessar
ycondi
ti
onf
oradeci
sion.i
.e:Ther
ewoul
dbenoneedf
or
deci
sionsi
fpr
obl
emsdi
dnotexi
st.
2.Det
ermi
net
hesetofal
ter
nat
ivesol
uti
ons.
3.Det
ermi
net
hecr
it
eri
atoeval
uat
eal
ter
nat
ives.
=>I
dent
if
yingt
hosechar
act
eri
sti
cst
hatar
eimpor
tanti
nmaki
ngt
hedeci
sion.
4.Anal
yzet
heal
ter
nat
ives.
==>Theadv
ant
agesanddi
sadv
ant
agesofeachal
ter
nat
ive.
5.Sel
ectt
hebestal
ter
nat
ive
==>Sel
ectt
hebestal
ter
nat
ivet
hatsui
tst
osol
veourdeci
sionpr
obl
em.I
nsel
ect
ing
t
hebestal
ter
nat
ive,f
act
orssuchasr
isk,t
imi
ngandl
i
iti
ngf
act
orsshoul
dbe
consi
der
edadequat
ely
6.I
mpementt
hesol
uti
on
==>Put
ti
ngt
hedeci
sioni
ntoact
ion
7.Est
abl
i
shi
ngacont
rolandeval
uat
ionsyst
em
==>Ongoi
ngact
ionsneedt
obemoni
tor
ed
==>Fol
l
owi
ngt
hedeci
sion
==>Ev
aluat
ether
esul
tsanddet
ermi
nei
fasat
isf
act
orysol
uti
onhasbeenobt
ained.

TheDeci
sion-
Maki
ngEnvi
ronment
Deci
si
onsar
emadeundert
hreebasi
ccondi
ti
ons:
 Deci
sionundercer
tai
nty
 Deci
sionunderr
isk
 Deci
sionunderuncer
tai
nty

i
.Deci
sionmaki
ngundercondi
ti
onsofcer
tai
nty
Thedeci
sionmakerhasper
fectknowl
edgeaboutt
heout
come.

2
I
nthi
ssi
tuat
ion,wear
ereasonabl
ysur
ewhatwi
l
lhappenwhenwemakeadeci
sion.
Thei
nfor
mat
ioni
sav
ail
abl
e&i
sconsi
der
edt
ober
eli
abl
e&weknow t
hecause&
ef
fectr
elat
ionshi
ps.
Exampl
e:I
fyoudeci
det
oinv
esty
ourmoneyi
nsav
ingaccounti
ntheCommer
cial
BankofEt
hiopi
a,Youar
ecer
tai
nthaty
ouwi
l
lear
nthr
eeper
centi
nter
est
.

i
i
.Deci
sionmaki
ngundercondi
ti
onofr
isk
Usual
l
y,deci
sionmaker
scannothav
eapr
eci
seknowl
edgeaboutt
heout
comeofa
deci
si
on.
Deci
si
onmaker
smayonl
ybeabl
etoat
tachapr
obabi
l
ityt
otheexpect
edout
comes
ofeachal
ter
nat
ive.
Undert
hissi
tuat
ion,
onemayhav
efact
ual
inf
ormat
ion,
buti
tmaybei
ncompl
ete.
Exampl
e:I
fwegambl
ebyt
ossi
ngaf
aircoi
n,t
hepr
obabi
l
ityt
hatat
ailwi
l
ltur
nupi
s
50%.

i
i
i.Deci
sionmaki
ngundercondi
ti
onsofuncer
tai
nty
I
tisacasewher
enei
thert
her
eiscompl
etedat
anorpr
obabi
l
iti
escanbeassi
gnedt
o
t
hesur
roundi
ngcondi
ti
on.I
tist
hemostdi
ff
icul
tforamanager
.
Some condi
ti
ons t
hatar
e uncont
rol
l
abl
e by managementi
ncl
ude compet
it
ion,
gov
ernmentr
egul
ati
ons,t
echnol
ogi
caladv
ances,t
heov
eral
leconomy
,&t
hesoci
al&
cul
tur
alt
endenci
esofsoci
ety
.

Exampl
e: A cor
por
ati
ont
hatdeci
dest
o expand i
tsoper
ati
on,l
aunchi
ng anew
pr
oduct
,ordev
elopi
ngofanewt
echnol
ogyi
nast
rongcount
rymayknowl
i
ttl
eabout
i
tscul
tur
e,l
aws,
economi
cenv
ironment
,orpol
i
tics.Thepol
i
ticalsi
tuat
ionmaybeso
v
olat
il
ethatev
enexper
tscannotpr
edi
ctapossi
blechangei
ngov
ernment
.

1.
3.QUNATI
TIATI
VEANALYSI
SPROCESS
*
Qual
i
tat
iveski
l
lcanbedev
elopedt
hroughexper
ience.I
tisi
nher
ent
.
Thet
hreebasi
cquant
it
ati
veanal
ysi
spr
ocessesar
e:
1. Modeldevelopment
2. Datapreparat
ion
3.Modelsolut
ion

1.Modeldevel
opment
Modeli
sar
epr
esent
ati
onofr
ealobj
ect
s/si
tuat
ions

Fol
l
owi
ngar
ety
pesofmodel
s:
a.Physi
cal(
icon)model

3
Thi
sist
her
epr
esent
ati
onoft
hesi
tuat
ion,
probl
em orobj
ect
.
I
tis al ed St
so cal
l ati
c Model
.Iti
s gi
ven i
ntwo ort
hree di
mensi
ons.I
tisa
r
epr
esent
ati
onoft
her
eal
obj
ect
.
Exampl
e:
 Thestructur
eofanat om
 Modelofanai rpl
ane
 Photographofamachi ne
 Layoutdrawingofafactor
y
 Glob
b.Anal
ogueModel
s:
Thesear
eabst
ractmodel
smost
lyshowi
ngi
nterandi
ntr
arel
ati
onshi
psbet
weent
wo
ormor
epar
amet
ers.
Forexampl
e:
I
tmayshowt
her
elat
ionshi
pbet
weenani
ndependentv
ari
abl
e(i
nput
)wi
tht
hatofa
dependentv
ari
abl
e(out
put
).Fori
nst
ance;
hist
ogr
am,
frequencyt
abl
e,cause-
eff
ect
di
agr
am,
flowchar
ts,
Gant
tchar
ts,
pri
ce-
demandgr
aph,
wor
ldmapandot
her
s.
:
Ist
wodi
mensi
onal
.

c.Mat
hemat
icalmodel
s
Thi
sisal
soanabst
ractmodel
.Her
easetofr
elat
ionsi
srepr
esent
edi
nthef
orm of
mat
hemat
ical
equat
ions,
usi
ngsy
mbol
stor
epr
esentv
ari
ouspar
amet
ers.
Exampl
e:
2 2 2
x+y)
1.( =x+2xy
+y

Z=3000x1+2500x2
2.Max.
Subj
ectt
o:
2x1+x2<40
x1+3x2<45
x1 <12
x1,x2>0

x1andx2 ar
edeci
sionv
ari
abl
es

Obj
ect
iveFunct
ion:
Amat
hemat
ical
stat
ementoft
hegoal
ofanor
gani
zat
ion,
stat
edasi
ntentt
o
maxi
mizeort
omi
nimi
zesomei
mpor
tantquant
it
ysuchaspr
ofi
tsorcost
s.
.Z=3000x1+2500x2i
Max stheobj
ect
ivef
unct
ion
Const
rai
nt:
Ar
est
ri
cti
onont
her
esour
cesav
ail
abl
etoaf
ir
m(st
atedi
nthef
orm ofani
nequal
i
tyor
anequat
ion.
)
2x1+x2<40
Resource
const
raint
s
4
 x1+3x2<45 Ar
econst
rai
nts
x1 <12
x1,x2>0 x1,
x2>0i
snon-
negat
ivi
tyconst
rai
nt

2.Dat
apr
esent
ati
on
Dat
arepr
esentt
hev
aluesofi
nput
stot
hemodel
.
Max.Z=3000x1+2500x2
Subj
ectt
o:
2x1+x2<40
x1+3x2<45
x1 <12
x1,x2>0

3.Modelsol
uti
on
:
Isopt
imal
sol
uti
on
Fr
om t
heabov
eequat
i x1=12,
on, x2=11andMax.Z=$6350

Tr
ialander
rormet
hod
I
nthi
smet
hodacer
tai
nal
gor
it
hm i
sdev
eloped.Onest
art
ingpoi
nti
sani
nit
ial
sol
uti
on,whi
chi
sthef
ir
stappr
oxi
mat
ion.
Themet
hodofsol
uti
on,i
srepeat
edwi
tha
cer
tai
nsetofr
ulessot
hati
nit
ialsol
uti
oni
sgr
adual
l
ymodi
fi
edateachsubsequent
sol
uti
on;t
il
lopt
imalsol
uti
oni
sreached.Ther
ear
ecer
tai
ncr
it
eri
alai
ddownt
ocheck
whet
hert
hesol
uti
onhasbecomeanopt
imalsol
uti
on.Theset
ri
alsol
uti
onsar
ecal
l
ed
I
ter
ati
onsandt
hemet
hodi edi
scal
l ter
ati
vepr
ocess.Thecl
assi
calexampl
eoft
ri
al
ander
rormet
hodi
sLi
nearPr
ogr
ammi
ng
Thent
henextst
epaf
teri
ter
ati s model
oni test
ingandval
i
dat
ion.

1.
4.MANAGEMENTSCI
ENCEI
NPRACTI
CE
Ther
ear
evar
ioust
echni
quesusedi
nO.
R.Someoft
hesear
eli
stedher
e:
o Probabil
ityt
heor y
o Li
nearPr ogrammi ng
o Transportat
ionalgorit
hm
o Assignmentpr oblems
o QueuingTheor y
o PERT/ CPM Met hodetc
OR t
echni e’
quesar applicati
onspeci fic’
.Maxi
mum benef
itcanbeder
ived f
rom
sel
ect
ingmostappr
opr
iat
etechni
quesf
oreachspeci
fi
car
eaorpr
obl
em.Appr
opr
iat
e
sel
ect
ionofORi
sanequal
l
yimpor
tantt
ask.Eacht
echni
quehasi
tsownadv
ant
ages
andl
i
mit
ati
ons.
Inal
lsuchcases,t
heabi
l
ityoft
heManageri
stest
edi
nappr
opr
iat
e

5
sel
ect
ionofORt
echni
que.

CHAPTERI
I
LI
NEARPROGRAMMI
NG

Li
nearPr
ogr
ammi
ngi
samat
hemat
icalpr
ocesst
hathasbeedev
elopedt
ohel
p
managementi
ndeci
sionmaki
ngi
nvol
vi
ngt
heef
fi
ciental
l
ocat
ionofscar
esr
esour
ces
t
oachi
eveacer
tai
nobj
ect
ive.

Di
agr
ammat
ical
ly,

Scares Tobeal
locat
edt
o:
Resource
6

Resour
ce
LPi
samet
hodf
orchoosi
ngt
hebestal
ter
nat
ivef
rom asetoff
easi
bleal
ter
nat
ives
Toappl
yLP,
thef
oll
owi
ngcondi
ti
onsmustbesat
isf
ied:
a.Obj
ect
iveFunct
ion
I
sthegoalorobj
ect
iveofamanagement
,st
ated asi
ntentt
o maxi
mizeort
o
mi
nimi
zesomei
mpor
tantquant
it
ysuchaspr
ofi
tsorcost
s?
b.Const
rai
nts
Const
rai
nts Ar
eli
mit
ati
ons or r
est
ri
cti
ons i
mposed by t
he pr
obl
ems.And
const
rai
ntsi
ncl
ude:
1.Resour
ceconst
rai
nts:Ar
erest
ri
cti
onst
hatshoul
dbecl
ear
lyi
dent
if
iabl
eand
measur
abl
einquant
it
ati
vet
erms,
whi
char
isef
rom l
i
mit
ati
onofav
ail
abl
eresour
ces.

Exampl
esofl
i
mit
edr
esour
ces:
 Pl
antcapaci
ty
 Rawmat
eri
alsavai
l
abi
l
ity
 Laborpower
 Mar
ketdemand,
etc
2.Non-
negat
ivi
tyconst
rai
nts:
:
Areconst
rai
ntst
hatr
equi
ret
hedeci
sionv
ari
abl
esnott
otakeonnegat
ivev
alues?
c.Li
near
it
y
TheObj
ect
iveFunct
ionandt
heconst
rai
ntsmustbel
i
neari
nnat
urei
nor
dert
ohav
ea
Li
nearPr
ogr
ammi
ngPr
obl
ems(
LPP)

7
d.Feasi
bleal
ter
nat
ive
Ther
eshoul
dbeaser
iesoff
easi
bleal
ter
nat
ivecour
seofact
ionav
ail
abl
etot
he
deci
si
on-
maki
ngdet
ermi
nedbyr
esour
ceconst
rai
nts.Thus,wehav
etochooset
he
bestal
ter
nat
ive
Li
nearPr
ogr
ammi
ngPr
obl
emscanbesol
vedbyusi
ng:
i
.TheGeomet
ri
cmet
hodcal
l
ed”Gr
aphi
calMet
hod”
i
i
.TheAl
gebr
aicmet
hodcal
l
ed”Si
mpl
exMet
hod”

2.
1.FORMULATI
ONOFLP
Deci
si
onv
ari
abl
esar
ethev
ari
abl
eswhosev
aluesar
eunknownandar
esear
ched
Thecoef
fi
cient
soft
hev
ari
abl
esi
ntheObj
ect
iveFunct
ionar
ecal
l
edt
hepr
ofi
torcost
coef
fi
cient
s.Theyexpr
esst
her
ateatwhi
cht
hev
alueoft
heObj
ect
iveFunct
ion
i
ncr
easesordecr
easesbyi
ncl
udi
ngi
nthesol
uti
ononeuni
tofeachoft
hedeci
sion
v
ari
abl
es.
Thecoef
fi
cient
soft
heconst
rai
nts’v
ari
abl
esar
ecal
l
edt
hei
nput
-out
putcoef
fi
cient
s
t
hati
ndi
cat
ether
ateatwhi
cht
hegi
venr
esour
cesar
edepl
etedorut
il
ized.
Exampl
e:

2.
2.GRAPHI
CALSOLUTI
ON
Touset
hegr
aphi
cmet
hod,
thef
oll
owi
ngst
epsar
eneeded:
1.I
dent
if
ythepr
obl
em
i
.e.
:Thedeci
sionv
ari
abl
es,
theobj
ect
ivef
unct
ionandt
heconst
rai
nts
2.Dr
awagraphi
ncl
udi
ngal
ltheconst
rai
ntsandident
ifythef
easi
bleregion
3.Obtai
napointonthefeasi
bleregi
ont hatopti
mi zest
heobj ect
ivefunct
ion-
Opt
imalsol
uti
on
4.I
nter
pri
tet
her
esul
ts

 Gr
aphi
calLPi
sat
wo-
dimensi
onalmodel
.
A Maxi
mizat
ionPr
obl
em
==>Maxi
mizeZwi
thi
nequal
i
tiesofconst
rai
nt n<f
si orm

8
Exampl
e:Consi
dert
womodel
sofcol s;ModelAandB,ar
orTVset epr
oducedbya
companyt
omaxi
mizepr
ofi
t.Thepr
ofi
treal
i
zedi
s$300f
rom Aand$250f
rom setB.
Thel
i
mit
ati
onsar
e
A.av
ail
abi
l
ityofonl
y40hr
sofl
aboreachdayi
nthepr
oduct
iondepar
tment
.
B.adai
l
yav
ail
abi
l
ityofonl
y45hr
sonmachi
net
ime
C.abi
l
ityt
osal
e12setofmodel
A.

Howmanyset
sofeachmodelwi
l
lbepr
oducedeachdaysot
hatt
het
otalpr
ofi
twi
l
l
beasl
argeaspossi
ble?

Resour
cesusedperuni
t
Const
rai
nts ModelA ModelB Maxi
mum Avai
labl
ehr
s.
(
X1) (X2)

Laborhr
. 2 1 40

Machi
nehr
. 1 3 45
Mar
ket
inghr
. 1 0 12

Pr
ofi
t $300 $250

Sol
uti
on
1.For
mul
ati
onofmat
hemat
ical
model
i
ngofLPP
MaxZ=300X1+250X2
St
:
2X1+X2<40
X1+3X2<45
LPP
X1 <12
Model
X1,X2 >0
2.Conv
ertconst
rai
ntsi
nequal
i
tiesi
ntoequal
i
ties
2X1+X2=40
X1+3X2=45
X1 =12
3.Dr
awt
hegr
aphbyi
nter
cept
s
2X1+X2=40==>(
0,40)and(
20,
0)
X1+3X2=45==>(
0,15)and(
45,
0)

9
 X1 =12==>(
12,
0)
X1,X2 =0

2X1+X2=40
X2
X1=0
40 X1=12

15B

X1+X2=45
Feasi e C(
bl 12,
Region 11)
X2=0
X1
D
A 12 20 45
4.I
dent
if
ythef
easi
blear
eaoft
hesol
uti
onwhi
chsat
isf
iesal
lconst
rai
ns.
5.I
dent
if
ythecor
nerpoi
ntsi
nthef
easi
bler
egi
on
A(
0,0)
,B(
0,15)
,C(
12,
11)andD(
12,
0)
6.I
dent
if
ytheopt
imal
poi
nt
7.I
nter
pret
ether
esul
t

Cor
ner
s Coor
dinat
es MaxZ=300X1+250X2
A (
0,0) $0
B (
0,15) $3750
C (
12,
11) $6350
D (
12,
0) $3600

I
nter
pret
ati
on:
12uni
tsofpr
oductAand11uni
tsofpr
oductBshoul
dbepr
oducedsot
hatt
het
otal
pr
ofi
twi
l
lbe$6350.
Exer
cise:
1.A manuf
act
urerofl
i
ghtwei
ghtmount
aint
ent
smakest
wot
ypesoft
ent
s,
REGULAR t
entand SUPER t .EachREGULAR t
ent entr
equi
res1l
abor
-hour
f
rom t
hecut
ti
ngdepar
tmentand3l
abor
-hour
sfr
om t
heassembl
ydepar
tment
.
EachSUPER t
entr
equi
res2l
abor
-hour
sfr
om t
hecut
ti
ngdepar
tmentand4
l
abor
-hour
sfr
om t
he assembl
y depar
tment.
The maxi
mum l
aborhour
s
av
ail
abl
eperweeki
nthecut
ti
ngdepar
tmentandt
heassembl
ydepar
tment
ar
e32and84r
espect
ivel
y.Mor
eov
er,
thedi
str
ibut
or, wi
becauseofdemand, l
l
nott
akemorethan12SUPERtentsperweek.Themanuf
act
urersal
eseach
REGULARt
entsfor$160andcost
s$110pertentt
omake.WhereasSUPER
t
ental
esf
or$210pert
entandcost
s$130pert
entt
omake.

10
Requi
red:
A.For
mul
atet
hemat
hemat
ical
model
oft
hepr
obl
em
B.Usi
ngt
hegr
aphi
cmet
hod,det
ermi
nehowmanyofeacht
entt
hecompanyshoul
d
manuf
act
ureeacht
entt
hecompanyshoul
dmanuf
act
ureeachweeksoast
o
maxi
mizei
tspr
ofi
t?
C.Whati
sthi
smaxi
mum pr
ofi
tassumi
ngt
hatal
lthet
ent
smanuf
act
uredi
neach
weekar
esol
dint
hatweek?

Solut
ion
___
_ _
____
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
Laborhour
spertent
Depar
tment REGULAR(X1) SUPER(
X2) Maxi
mum l
abor
-hour
savai
labl
eper
week
_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
__

Cut
ti
ngdepar
tment 1 2 32
Assembl
ydepar
tment 3 4 84
Sell
ingpr i
cepert
ent $160 $210
Costpert ent $110 $130
Profitpertent $50 $80

*
The di
str
ibut
orwi
l
lnott
ake mor
ethan 12 SUPER t
ent
s perweek.Thus,t
he
manuf
act
urershoul
dnotpr
oducemor
ethan12SUPERt
ent
sperweek.
LetX1=TheNoofREGULARt
ent
spr
oducedperweek.
X2=TheNoofSUPERtentspr
oducedperweek.
X1andX2ar
ecall
edthedeci
sionv
ari
ables

……….
Cutt
ingdepar
tmentconst
rai
nt
LPPModel
……….
Assemblydepar
tmentconst
rai
nt
……….
Demandconst
rai
nt

……….
Non-
negat
ivi
tyconst
rai
nts

Cor
ner
s Coor
dinat
es MaxZ=50X1+800X2

11
A (
0,0) $0
B (
0,12) $960
C (
8,12) $1360
D (
20,
6) $1480
E (
28,
0) $1400

I
nter
pret
ati
on:
Themanuf
act
urershoul
dpr e20REGULARt
oduceandsal sand6SUPERSt
ent ent
sto
getamaxi
mum weekl
ypr
ofi
tof$1480.

B Mi
nimi
zat
ionPr
obl
em
==>Mi
nimi
zeZwi
thi
nequal
i
tiesofconst
rai
nt n>f
si orm
Exampl
e:
1.Supposet
hatamachi
neshophast
wodi
ff
erentt
ypesofmachi
nes;
machi
ne1
andmachi
ne2,
whi
chcanbeusedt
omakeasi
ngl
epr
oduct.
Thesemachi
nes
v
aryi
ntheamountofpr
oductpr
oducedperhr
.,i
ntheamountofl
aborused
andi
nthecostofoper
ati
on.
Assumet
hatatl
east acer
tai
namountofpr
oductmustbepr
oducedandt
hatwe
woul
dli
ket
out
il
izeatl
eastt
her
egul
arl
aborf
orce.Howmuchshoul
dweut
il
izeeach
machi
nei
nor
dert
out
il
izet
otal
cost
sandst
il
lmeet
sther
equi
rement
?
Solut
ion
___
_ _
____
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
_
Resour
ceused

Machi
ne1(
X1) Machi
ne(
X2) Mi
nimum r
equi
redhour
s
_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
__

Pr
oductpr
oduced/
hr 20 15 100
Labor
/hr 2 3 15_
___
___
_
Oper
ati
onCost $25 $30_
___
___
___
___
___
___
___
___
___
_

LPPModel

Const
rai
ntequat
ion:

12
 20X1+15X2=100 ==>(
0,20/
3)and(
5,0)
2X1+3X2=15 ==>(
0,5)and(
7.5,
0)
X1X2>0
X2
X1=0
A(
0,20/
3)

Feasi
bleRegi
on

B(
2.5,
3.
33) X2=0
X1
5 C(
7.5,
0)

__
______
___
___
___
___
___
___
___
___
___
___
___
___
_____
___
___
___
___
___
___
___
___
____
____
___
___
____
_ _
_
Corner
s Coor
dinat
es MinZ=25X1+30X2
A (
0,20/
3) 200

B (
2.5,
3.33) 162.
5
C (
7.5,
0) 187.
5
_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
__
X1=2.5
X2=3.
33and
MinZ=162.
5

Exer
cise:
1.Acompanyownst
wof
lourmi
l
ls(
AandB)whi
chhav
edi
ff
erentpr
oduct
ion
capaci
ti orHI
es f GH,MEDI
UM and LOW gr
ade f
lour
.Thi
s companyhas
ent
eredcont
ractsuppl
yfl
ourt
oaf
ir
m ev
eryweekwi
th12,8,and24qui
ntal
s
ofHI
GH,MEDI
UM andLOW gr
ader
espect
ivel
y.I
tcost
stheCo.$1000and
$800perdayt
orunmi
l
lAandmi
l
lBr
espect
ivel
y.Onaday
,mi
l
lApr
oduces6,
2,
and4qui
nt sofHI
al GH,
MEDI
UM andLOW gr
adef
lourr
espect
ivel
y.
Mi
l
lB pr
oduces2,2 and 12 qui
nt sofHI
al GH,MEDI
UM and LOW gr
ade f
lour
r
espect
ivel
y.Howmanyday
sperweekshoul
deachmi
l
lbeoper
atedi
nor
dert
omeet
t
hecont
ractor
dermosteconomi
cal
l
yst
andar
dize?Sol
vegr
aphi
cal
l
y.
Sol
uti
on:
Noofday
sperweekof Mi
nimum f
louri
n
Mi
l X1)
A( Mi
l
l X2)
B( qui
ntal
s
HI
GHCapacity(
inqui
ntal
) 6 2 12
MEDI
UM Capaci
ty(i
nquintal
) 2 2 8
LOW Capaci
ty(
inqui
ntal
) 4 12 24

13
 $1000 $800

Const
rai
ntequat
ion:

(
0,6)
,(2,
0)
(
0,4)
,(4,
0)
(
0,2)
,(6,
0)

Cor
ner
s Mi
nZ=$1000X1+800X2
(
0,6) $4800
(
1,3) $3400
(
3,1) $3800
(
6,0) $6000
X1=1
X2=3and
MinZ=$3400

14
 X2
X1=0
6 6X1+2X2=12
2X1+2X2=8
4 FR
4X1+12X2=24
(
1,3)
(
3,1)
X2=0
X1
2 4 6

Not
e:
-
Inmaxi
mizat
ionpr
obl
ems,
ourpoi
ntofi
nter
esti
slooki
ngt
hef
urt
hestpoi
ntf
rom t
he
or
igi
n.
-
Inmi
nimi
zat
ionpr
obl
ems,ourpoi
ntofi
nter
esti
slooki
ngt
hepoi
ntnear
estt
othe
or
igi
n.

2.
3.SPECI
ALCASESI
NGRAPHI
CSMETHODS
1.RedundantConst
rai
nt
I
faconst
rai
ntwhenpl
ott
edonagr
aphdoesn’
tfor
m par
toft
heboundar
ymaki
ng
t
hef
easi
bler
egi
onoft
hepr
obl
em t
hatconst
rai
nti
ssai
dtober
edundant
.

Exampl
e:
Af
ir
misengagedi
npr
oduci
ngt
wopr
oduct
sAandB.
Eachuni
tofpr
oductA
r
equi
res2Kgofr
awmat
eri
aland4l
abor
-hr
s.f
orpr
ocessi
ng.Wher
easeachuni
t
ofpr
oductBr
equi
res3Kgofr
awmat
eri
alsand3hr
sofl
abor
.Ev
eryuni
tofpr
oduct
Aneeds4hr
stopackagi
ngandev
eryuni
tofpr
oductBneeds3.
5hr
sforpackagi
ng.
Ev
eryweekt
hef
ir
m hasav
ail
abi
l
ityof60Kgofr
aw mat
eri
al,96l
abor
-hour
sand
105hr
sIt
hepackagi
ngdepar
tment
.1uni
tofpr
oductAsol
dyi
elds$40pr
ofi
tand
1uni
tofBsody
iel
ds$35pr
ofi
t.
Requi
red:
a.For
mul
atet
hispr
obl
em asaLPP
b.Fi
ndt
heopt
imal
sol
uti
on

Sol
uti
on
_
____
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
_

15
 Pr
oduct
s Resour
ceavai
labl
e
Resour
ces A B perweek
_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
__
Rawmat
eri
als(
Kg) 2 3 60
Labor(
hr) 4 3 96
Packagi
ng(
hr) 4 3.
5 105
Pr
ofi
tperuni
t $40 $35

LetX1=TheNoofuni
tsofpr
oductApr
oducedperweek
o
X2=TheN ofuni
tsofpr
oductBpr
oducedperweek
a.LPPModel

X2
(
0,32)

Labor
:4X1+3X2=96

(
0,30)
Packagi 4X1+3.
ng: 5X2=105
(
0,20) C(
18,
8)
Rawmat
eri
al2X1+3X2=60
:
FR
X1
A(
0,0) D(
24,
0) (
26,
0) (
30,
0)

 Thepackagi
nghri
sredundant
.

Cor
ner
s Coor
dinat
es Mi
nZ=40X1+35X2
A (
0,0) 0
B (
0,20) 700
C (
18,
8) 1000
D (
24,
0) 960
X1=18
X2=8and

16
 nZ=1000
Mi

I
nter
pret
ati
on:
Thecompanyshoul
dpr
oduceandsal
e18uni
tsofpr
oductAand8uni
tsofpr
oductB
perweeksoast
ogetamaxi
mum pr
ofi
tof1000.
 Byt
hispr
oduct
ionpl
ant
heent
ir
erawmat
eri
alwi
l
lbeconsumed.
 2X1+3X2<60
2(
18)+3(
8)=60
60=60==>Noi
dleorunusedr
awmat
eri
al
 4X1+3X2<96
4(
18)+3(
8)<96
96=96==>t
heent
ir
elaborhourwi
l
lbeconsumed
 4X1+3.
5X2<105
100<105==>Ther
eist
o bei
dleorunusedcapaci
tyof5hr
sint
hepackagi
ng
depar
tment
.
Not
e:
Thepackagi
nghour
’sconst
rai
ntdoesnotf
orm par
toft
heboundar
ymaki
ngt
he
f
easi
bler
egi
on.Thus,t
his const
rai
nti
s ofno consequence and i
sther
efor
e,
r
edundant
.Thei
ncl
usi
onorexcl
usi
onofar
edundantconst
rai
ntdoesnotaf
fectt
he
opt
imalsol
uti
onoft
hepr
obl
em.

2.Mult
ipl
eopti
malSoluti
ons
/Al
ter
nati
veopti
malsolut
ions/
-
Thi
sisasi
tuat
ionwher
ebyaLPPhasmor
ethanoneopt
imal
sol
uti
on.
Mul
ti
pleopt
imalSol
uti
onswi
l
lbef
oundi
ftwocor
ner
sgi
veopt
imalsol
uti
on,
thent
he
l
i
nesegmentj
oini
ngt
hesepoi
ntswi
l
lbet
hesol
uti
on.
==>Wehav
eunl
i
mit
ednumberofopt
imalsol
uti
onwi
thouti
ncr
easi
ngordecr
easi
ng
t
heobj
ect
ivef
unct
ion.

Exampl
e:
Thei
nfor
mat
iongi
venbel
owi
sfort
hepr
oduct
sAandB.
_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
__
Machi
nehour
sperweek Maximum avai
labl
e
Depar
tment Pr
oductA Pr oductB perweek
_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
__

Cut
ti
ng 3 6 900
Assembl
y 1 1 200
Pr
ofi
tperuni
t $8 $16
__
_____
___
___
____
___
___
____
___
___
___
___
_ __
____
___
___
____
___
___
____
___
Assumethatthecompanyhasamar keti
ngconst
rai
ntonsell
i
ngpr oduct
sBand

17
 t
her
efor
eitcansal
eamaxi
mum of125uni
tsoft
hispr
oduct
.
Requi
red:
a.For
mul
atet
heLPPoft
hispr
obl
em
b.Fi
ndt
heopt
imal
sol
uti
on
Sol
uti
on:
LetX1=TheNoofuni
tsfpr
oductAproducedperweek
o
X2=TheN ofuni
tsfpr
oductBproducedperweek
a.TheLPPModel oft
heprobl
em is:

X2 X1=0
(
0,200)

(0,
150) C(
50,
125) X2=125Mar
ket
ingequat
ion
B(
0,125)
D(
100,
100)
Cut
ti
ng:3X1+6X2=900
FR X2=0

A(
0,0) X1
E(
200,
0) (
300,
0)

Cor
ner
s Coor
dinat
es MaxZ=8X1+16X2
A (
0,0) 0
B (
0,125) 2000
C (
50,125) 2400
D (
100,100) 2400
E (
200,
100) 1600

I
nter
pret
ati
on:
hCandDar
Bot eopt
imalsol
uti
ons.Anypoi
ntont nesegmentCDwi
hel
i l
lal
sol
eadt
o
t
hesameopt
imal
sol
uti
on.
==>Mul
ti
pleopt
imalsol
uti
onspr
ovi
demor
echoi
cesf
ormanagementt
oreacht
hei
r
obj
ect
ives.

18
 3.I
nfeasi
bleSol
uti
on
Asol
uti
oni
scal
l
edf
easi
blei
fitsat
isf
iesal
ltheconst
rai
ntsandt
heconst
rai
ntsand
non-
negat
ivi
tycondi
ti
on.Howev
er,i
tissomet
imespossi
blet
hatt
heconst
rai
ntsmay
bei
nconsi
stentsot
hatt
her
eisnof
easi
blesol
uti
ont
othepr
obl
em.Suchasi
tuat
ion
i edi
scal
l nfeasi
bil
it
y.

Exampl
e:
MaxZ=20X1+30X2
St
:
2X1+X2<40
4X1+X2<60
X1 >30
X1,X2 >0
Sol
uti
on:

X2 X1=0
(
0,60) X1=30
4X1+X2=60
(0,
40)

2X1+X2=40
X2=0
X1
(
15,
0) (
20,
0) (
30,
0)

Not
e:
-
Int
heabov
egr
aph,
ther
eisnocommonpoi
nti
ntheshadedar
ea.
-
All
const
rai
ntscannotbesat
isf
iedsi
mul
taneousl
yandt
her
eisnof
easi
blesol
uti
ont
o
t
hepr
obl
em.

4.Mi
xofconst
rai
nts
Exampl
e:
ABC Gasol
i
ne Companyhas t
wo r
efi
ner
ies wi
th di
ff
erentpr
oduct
ion capaci
ti
es.
Ref
iner
yAcanpr
oduce4, onsperdayofSUPERUNLEADDGASOLI
000gal
l NE,2000
onsperdayofREGULARUNLEADED GASOLI
gal
l NE and1000gal
l
onsperdayof

19
LEADEDGASOLI NE.Ontheot
herhand,r
efi
ner
yBcanpr
oduce1000gal
lonsperday
ofSUPER UNLEADED,3000gallonsperdayofREGULAR UNLEADED and4,000
gal
lonsperdayofLEADED.

Thecompanyhasmadeacont
ractwi
thanaut
omobi
l
emanuf
act
urert
opr
ovi
de
nesofSUPERUNLEADED,42000gal
24000gasol
i onsofREGULARUNLEADEDand
l
onsofLEADED.
36000gal
l Theaut
omobi
l
emanuf
act
urerwant
sdel
i
ver
yinnotmor
e
t
han14day
s.

Thecostofr
unni
ngr
efi
ner
yAi
s$1500perdayandr
efi
ner
yBi
s$2400perday
.

Requi
red:
a.For
mul
atet
hispr
obl
em asaLPP
b.Det
ermi
net
henumberofday
sthegasol
i
necompanyshoul
doper
ateeachr
efi
ner
y
i
nor
dert
omeett
het
ermsoft
heabov
econt
ractmosteconomi
cal
.(
i.
e.Ata
mi
nimum r
unni
ngcost
)
c.Whi
chgr
adeofgasol
i
newoul
dbeov
erpr
oduced?

Sol
uti
on:
___
_____
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
_
Pr
oduct
ionperday(
ingal
lons) Cont
ractwi
thanaut
omobi
lemanuf
act
urer
Gr
adeofgasol
ine A B
_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
__

SUPERUNLEADED 4000 1000 24,

000
REGULARUNLEADED 2000 3000 42,
000
LEADED 1000 4000 36,
000
Runni
ngcostperday $1,
500 $2,
400
_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
__

 Theaut
omobi
l
emanuf
act
urerwant
sdel
i
ver
yinnotmor
ethan14day
s.
LetX1=TheNoofday
sref
iner
yAshoul
dwor
k.
X2=TheNoofdaysr
efi
ner
yBshoul
dwor
k.
a.LPPoft
hepr
obl
em

MinZ=1500X1+2400X2
St
:
4000X1+1000X2>24000
2000X1+3000X2>42000

20
 1000X1+2000X2>36000
X1 <14

X2<14
X1,
X2 >0
==>Tosi
mpl
i
fyt
hepr
obl
em di
vi
deby1000t
heconst
rai
nts

MinZ=1500X1+2400X2
St
:
4X1+1X2>24
2X1+3X2>42
X1+4X2 >36
X1 <14
X2<14
X1,
X2 >0
Del
i
ver
yti
me:
X1=14
4X1+X2=24
SUG:
24

A(
2.5,
14) B(
14,
14) Del
i
ver
yti
me:
X2=14
FR

D(
12,
6) C(14,
5.
5) LG:
X1+4X2=36

2X1+3X2=42
RUG:

(
6,0) (
14,
0) (
21,
0) (
36,
0)
Not
e:Poi
ntA,
B,C,
andDar
esol
vedbyel
i
minat
ion-
subst
it
uti
onmet
hod

_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
Cor
ner
s Coor
dinat
es MaxZ=1500X1+2400X2
A (
2.5,
4) $37350
B (
14,
14) 54600
C (
14,
5.5) 34200
D (
12,
6) 32400
E (
3,12) 33300
_
______
_______
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
I
nter
pretat
ion:
Theoi
lcompanyshoul
doper
ater
efi
ner
yAf
or12day
sandr
efi
ner
yBf
or6day
sata

21
mi
nimum oper
ati
ngcostof$32,
400.
c.I
sther
eanyov
erpr
oduct
ion
SUG: 4000X1+1000X2>24000
4000(12)+1000(6)>24000
54000>24000
Ther
efor
e,30,
000gallonsoverpr oduct
ion

RUG: 2000X1+3000X2>42000
2000(12)+3000(6)>42000
42000>42000
Ther
efor
e,ther
eisnooverpr oduct
ionofRUG

LG: 1000X1+4000X2>36000
1000(12)+1000(6)>36000
36000>36000
Ther
efore,NooverproductionofLG

5.UnboundedSol
uti
on
Whent
hev
alueofdeci
sionv
ari
abl
esi
nLPi
sper
mit
tedt
oincr
easei
nfi
nit
elywi
thout
v
iol
ati
ngt
hef
easi
bil
i
tycondi
ti
on,
thent
hesol
uti
oni
ssai
dtobeunbounded.
Her
e,t
he
obj
ect
ivef
unct
ionv
aluecanal
sobei
ncr
easedi
nfi
nit
ely
.Howev
er,anunbounded
f
easi
bler
egi
onmayy
iel
dsomedef
ini
tev
alueoft
heobj
ect
ivef
unct
ion.
Exampl
e:
Uset
hegr
aphi
cal
met
hodt
osol
vet
hef
oll
owi
ngLPP.
1.Max.
Z=3X1+4X2
St
:
X1-
X2<-1==>-
X1+X2>1si
ncet
hequant
it
ysol
uti
oni
sposi
ti
ve
-
X1+X2<0
X1,
X2 >0
X2 X1-
X2 =-
1

X1+X2 =0

1 Unbounded
Feasi
bleRegi
on

X1
22
 Fi
g:UnboundedSol
uti
on

2.Max.
Z=3X1+2X2
St
:
X1-X2<1
X1+X2<3
X1,X2 >0
X2

A(
0,3) Unbounded
Feasi
bleRegi
on
X1-
X2=1

B(
2,1)
X1+X2=3

X1

Not
eher
ethatt
het
wocor
ner
soft
her
egi
onar
eA(
0,3)and.
B(2,
1).
Thev
alueof
Max
Z(A)
=6andMaxZ(
B)=8.Butt
her
eexi
stnumberofpoi
ntsi
ntheshadedr
egi
onf
or
whi
cht
hev
alueoft
heobj
ect
ivef
unct
ioni
smor
ethan8.
Forexampl
e,t
hepoi
nt(
10,
12)
l
i
esi
nther
egi
onandt
hef
unct
ionv
alueatt
hispoi
nti
s70whi
chi
smor
ethan8.

Remar
k:
Anunboundedsol
uti
ondoesnotmeant
hatt
her
eisnosol
uti
ont
othegi
venLPP,
but
i
mpl
i
est
hatt
her
eexi
tsani
nfi
nit
enumberofsol
uti
ons.

Exer
cise:
Usegr
aphi
cal
met
hodt
osol
vet
hef
oll
owi
ngLPP.
1.Max.Z=7/4X1+3/2X2 2.Max.
Z=3X1+2X2
St: St
:
8X1+5X2<320 -2X1+3X2<9
4X1+5X2<20 X1-
5X2 >-20
X1 >15 X1,X2 >0
X2>10
X1,X2>0
Answer:Nof easi
blesolut
ion Answer
:Unboundedsol
uti
on

23
3.Max.Z=3X1+2X2 4.
Max. Z=X1+X2
St: St:
X1-X2<1 X1+X2<1
X1+X2>3 -
3X1+X2>3
X1,X2>2 X1,X2>0
Answer:Unboundedsol ut
ion Answer :Unboundedsol
uti
on
5.Max.Z=6X1-4X2 6.
Max. Z=X1+1/2X2
St: St:
2X1+4X2<4 3X1+3X2<12
4X1+8X2>16 5X1 <10
X1,X2 >0 X1 +X2 >8
Answer :I
nfeasi
blesolut
ion -
X1 +X2 >4
X1, X2 >0
Answer :Inf
easibl
esol
uti
on
Exer
cise
I
.Sol
vet
hef
oll
owi
ngLPpr
obl
emsusi
ngt
hegr
aphi
cal
met
hod.
1.Max.Z=15X1-10X2 2.Max.
Z=2X1+X2
St: St :
4X1+6X2<360 X1+2X2<10
3X1+0X2<180 X1+X2 <6
0X1+5X2<280 X1 -X2 <2
X1,X2 >0 X1 - 2X2<1
Answer :X1=60,X2 =20 X1,          X2  >0
andMax.Z=1,100 Answer:X1=4,X2 =2
andMax.Z=10

3.Max.Z=10X1+15X2 4.
Min.
Z=3X1+2X2
St: St
:
2X1+X2<26 5X1+X2>10
2X1+4X2<56 X1+X2 >6
-X1+X2<5 X1+4X2 >12
X1,X2 >0 X1,          X2  >0
Answer :X1=4,X2 =2 Answer :
X1=1,X2=5
andMax. Z=230 andMi n. Z=13
5.Mi
n.Z=- X1+2X2 6.
Min.
Z=20X1+10X2
St: St:
-X1+3X2<26 X1+2X2<40
X1+X2 <6 3X1+4X2>30

24
 X1-
X2<2 4X1+3X2>60
X1,X2 >0 X1,       X2  >0
Answer
:X1=2,X2 =0 Answer
:X1=6,X2=12 and
Min.
Z=-
2 andMin.Z=240

I
I.
Amanuf
act
urerpr
oducest
wodi
ff
erentmodel
s;XandY,oft
hesamepr
oduct.
The
r
awmat
eri
alsr
1andr
2ar
erequi
redf
orpr
oduct
ion.Atl
east18Kgofr
1and12Kgofr
2

mustbeuseddai
l
y.Al
mostatmost34hour
sofl
aborar
etobeut
il
ized.
2Kgofr
1ar
e
neededf
oreachmodelXand1Kgofr
1foreachmodelY.ForeachmodelofXandY,
1Kgofr
2isr
equi
red.I
ttakes3hour
stomanuf
act
ureamodelXand2hour
sto
manuf
act
ureamodelY.Thepr
ofi
treal
i
zedi
s$50peruni
tfr
om modelXand$30per
uni
tfr
om modelY.Howmanyuni
tsofeachmodelshoul
dbepr
oducedt
omaxi
mize
t
hepr
ofi
t?
Answer
:10uni
tsofmodelX,
2uni
tsofmodelYandt
hemaxi
mum pr
ofi
tis$560.

I
II
.A manuf
act
uri
ngf
ir
m pr
oducest
womachi
nepar
tsP1 andP2 usi
ngmi
l
li
ngand
gr
indi
ng machi
nes .
The di
ff
erentmachi
ning t
imes r
equi
red f
oreach par
t,t
he
machi
ningt
imesav
ail
abl
eondi
ff
erentmachi
nesandt
hepr
ofi
toneachmachi
nepar
t
ar
easgi
venbel
ow:
_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
_
Manuf
act
uri
ngt
ime Maxi
mum t
ime
r
equi
red(
min) av
ail
abl
eperweek(
min)
Machi
ne P1 P2
_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
__
Lathe 10 5 25,000
Milli
ngMachine 4 10 2000
GrindingMachine 1 1.5 450
Profitperuni
t($) $50 $100
_
____
___
___
____
____
____
___
__ __
___
___
____________
___
___
___
____
_____
___
Det
ermi
nethenumberofpiecesofP1 and P2 t
obemanufact
uredperweekt
o
maxi
mizepr
ofi
t.
Answer
:X1=375,X2 =50 andMax.
Z=23750

IV.Apersonrequir
es10, 12and12uni
tsofchemi sA,
cal BandCr especti
vel
yforhis
garden.Ali
quidproductcont
ains5,
2and1uni sofA,
t BandCrespect
ivel
yperjar
.A
dryproductcontai
ns1,2and4uni tsofA,BandCpercar ton.I
ftheli
quidproduct
sel
l
sfor$3perj
arandt
hedr
ypr
oductsel
l
s$2percar
ton,howmanyofeachshoul
d
bepur
chasedi
nor
dert
omi
nimi
zecostandmeett
her
equi
rement
?
Answer
:1Uni
tofl
iqui
d,5uni
tsofdr
ypr
oductandMi
n.Z=$13

25
 CHAPTERI
II
SI
MPLEXMETHOD

I
NTRODUCTI
ON
The gr
aphi
calmet
hod t
o sol
vng LPPspr
i ovi
desf
undament
alconcept
sforf
ull
y
under
standi
ngt
heLPpr
ocess.Howev
er,t
hegr
aphi
calmet
hodcanhandl
epr
obl
ems
i
nvol
vi
ngonl
ytwodeci
sionv
ari
abl sayX1andX2)
es( .

I sGeor
n1940’ geB.
Dant
zig devel
opedanal
gebr
aicappr
oachcal
l he Si
edt mpl
ex
Met
hod whi
ch i
s an ef
fi
cientappr
oach t
o sol
ve appl
i
ed pr
obl
ems cont
aini
ng
numer
ousconst
rai
ntsandi
nvol
vi
ngmanyv
ari
abl
est
hatcannotbesol
vedbyt
he
gr
aphi
cal
met
hod.

Thesi
mpl
exmet sanI
hodi TERATI
VEor“
stepbyst
ep”met
hodorr
epet
it
iveal
gebr
aic
appr
oacht
hatmov
esaut
omat
ical
l
yfr
om onebasi
cfeasi
blesol
uti
ont
oanot
herbasi
c
f
easi
blesol
uti
on i
mpr
ovi
ng t
hesi
tuat
ion each t
imeunt
ilt
heopt
imalsol
uti
on i
s
r
eachedat
.

Not
e:
Thesi
mpl
exmet
hodst
art
swi
thacor
nert
hati
sint
hesol
uti
onspaceorf
easi
ble
r
egi
onandmov
est
oanot
hercor
nerIt
hesol
uti
onspacei
mpr
ovi
ngt
hev
alueoft
he
obj
ect
ivef
unct
ioneacht
imeunt
ilopt
imal
sol
uti
oni
sreachedatt
heopt
imal
cor
ner
.

 Mat
ri
ces
Di
ff
erentt
ypesofmat
ri
ces:
1.Rowmat
ri
x
Amat
ri
xwhi
chhasexact
lyoner
ow
Exampl
e:
(
1234) ==>1x4mat
ri
xes
2.Col
umnmat
ri
x
Amat
ri
xwhi
chhasexact
lyonecol
umn
Exampl
e:
5
==>3x1Mat
ri
x
6

26
3.Squar
emat
ri
x
Amat
ri
xinwhi heNoofr
cht ows( heNoofcol
m)=t umns(
n)
Exampl
e:
1 2
==>2x2squar
eMat
ri
x
3 4

4.Nul
lorZer
oMat
ri
xes
-
Amat
ri
xeachofwhoseel
ement
siszer
o.
Exampl
e:

0 0
==>3x2Zer
oMat
ri
x
0 0

0 0
5.Di
agonalMat
ri
x
Theel
ement
saij(
i=j
)ar eddi
ecal
l agonalel
ement
sofasquar
emat
ri
x.
Exampl
e:

1 2 3
Thedi
agonal
element
sar
e:a11=1,
a22=5anda33=9
4 5 6

7 8 9

Asquar
emat
ri
xofev
eryel
ementot
hert
handi
agonal
element
siszer
o,i
scal
l
eda
di
agonalmat
ri
x.
Exampl
e:

1 0 0 0 0 0 0
0 5 0 0 5 and 0 0
0 0 9

6.I
dent
it
yMat
ri
x
Adi
agonal
mat
ri
xwhosedi
agonal
element
sar
eal
lequal
to1(
uni
ty)
Exampl
e:
1 0
I
sani
dent
it
ymat
ri
x
Sco
0peof
1sol
uti
onofLPPbysi
mpl
exmet
hod

27
Fol
l
owi
ngt
ypesofpr
obl
emsar
esol
vedbysi
mpl
exmet
hod:
 Maxi
mizeZwi
thi
nequal
i
tiesofconst
rai
ntsi
n“<“
for
m.

 Mi
nimi
zeZwi
thi
nequal
i
tiesofconst
rai
ntsi
n“>“
for
m.

 Maxi
mizeZwi
thi
nequal
i
tiesofconst
rai
ntsi
n“>“
for
m.

 Maxi
mizeZorMi
nimi
zeZwi
thi
nequal
i
tiesofconst
rai
ntsi
n“<“
,“>“
or“
=”f
orm.

3.
1.MAXI
MIZATI
ONPROBLEMS
 Max
imi
zeZwi
thi
nequal
i
tiesofconst
rai
ntsi
n“<“
for
m

Exampl
e:
Sol
vet
hepr
obl
em usi
ngt
hesi
mpl
exappr
oach
Max.Z=300x1+250x2
Subj
ectt
o:
2x1+x2<40(
Labor)
x1+3x2<45 (
Machi
ne)
x1 <12 ( Mar
ket
ing)
x1,x2 >0
Solut
ion
Step1
For
mul
ateLPPModel

St
ep2
St
andar
dizet
hepr
obl
em
I
.e.Conv
ertconst
rai
nti
nequal
i
tyi
ntoequal
i
tyf
orm byi
ntr
oduci
ngav
ari
abl
ecal
l
ed
Sl
ackvar
i e.
abl

Sl
ackVar
iabl
es:
A sl
ackv
ari
abl
e(s)i
saddedt
othel
efthandsi
deofa<const
rai
ntt
ocov
ertt
he
const
rai
nti
nequal
i
tyi
ntoequal
i
ty.Thev
alueoft
hesl
ackv
ari
abl
eshowsunused
r
esour
ce.
Asl
akev
ari
abl
eemer
geswhent
heLPPi
samaxi
mizat
ionpr
obl
em.

28
Sl
ackv
ari
abl
esr
epr
esentunusedr
esour
ceori
dlecapaci
ty.Thus,t
heydon’
tpr
oduce
anypr
oductandt
hei
rcont
ri
but
iont
opr
ofi
tiszer
o.
Sl
ackv
ari
abl
esar
eaddedt
otheobj
ect
ivef
unct
ionwi
thzer
ocoef
fi
cient
s.
hats1,
Lett s2,
ands3ar
eunusedl
abor
,machi
neandmar
ket
inghr
srespect
ivel
y.

Max.
Z=300x1+250x2+0s1+0s2+0s3
St
:
2x1+x2+s1+0s2+0s3=40 St
andar
dfor
m
x1+3x2+0s1+s2+0s3=45
x1+ 0s1+0s2+ s3=12
x1 ,x2 , s1,s2, s3 >0
St
ep3
Obt
aint
hei
nit
ial
simpl
ext
abl
eau
Tor
epr
esentt
hedat
a,t
hesi
mpl
exmet
hodusesat
abl
ecal
l hesi
edt mpl
ext
abl
eor
hesi
t mpl
exmat
ri
x.
==>I
nconst
ruct
ingt
hei
nit
ialsi
mpl
ext
abl
eau,t
hesear
chf
oroft
heopt
imalsol
uti
on
begi
nsatt
heor
igi
n.I
ndi
cat
ingt
hatnot
hingcanbepr
oduced;
Thus,
fir
stassumpt
ion,
Nopr
oduct
ioni
mpl
i hatx1=0andx2=0
est
==>2x1+x2+s1+0s2+0s3=40 ==>x1+3x2+0s1+s2+0s3=45
0)+0+s1+0s2+0s3=40
2( 0)+0s1+s2+0s3=45
0+3(
s1=40–Unusedl
aborhr
s. s2=45–Unusedmachi
nehr
s.

==>x1+0s1+0s2+s3=12
0+0s1+0s2+s3=12
s3=12–UnusedMar
ket
inghr
s.
Ther
efor
e,Max.
Z=300x1+250x2+0s1+0s2+0s3
=300(
0)+250(
0)+0(
40)+0(
45)+0(
12)
=0

Not
e:
I
ngener
al,whenev
ert
her en v
ear ari esand m const
abl rai
nts(
excl
udi
ngt
henon-
negat
ivi
ty) em i
,wher sl hann(
esst m<n)
,n-
m var
i esmustbesetequalt
abl ozer
o
bef
oret
hesol
uti
oncanbesol
vedal
gebr
aical
l
y.
a.Basi
cvar
iabl
esar
evar
iabl
eswi
thnon-
zer
osol
uti
onv
alues.
Or
:basi
cvar
iabl
esar
evar
iabl
est
hatar
eint
hebasi
csol
uti
on.Basi
cvar
iabl
es
e0v
hav aluesi
ntheCj-
Zjr
ow.
b.Non-
basi
cvar
iabl
esar
evar
iabl
eswi
thzer
osol
uti
onv
alues.

29
 Or
:non-
basi
cvar
iabl
esar
evar
iabl
est
hatar
eoutoft
hesol
uti
on.
==>n=5v
ari
abl x1 ,x2,s1,s2,ands3)andm=3const
es( rai
nts(
Labor
,machi
neand
mar
ket
ingconst
rai
nts)
,excl
udi
ngnon-
negat
ivi
ty.
[

Ther
ef e,n-
or m=5-
3=2v
ari
abl x1 andx2)ar
es( esetequalt
ozer
oi he1st si
nt mpl
ex
t
abl
eau.Thesear
enon-
basi
cvar
iabl
es.3Var
iabl s1,s2,ands3)ar
es( ebasi
cvar
iabl
es
(
i he1stsi
nt mpl
ext
abl
eau)becauset
heyhav
enon-
zer
osol
uti
onv
alues.
St
ep4
Const
ructt
hei
nit
ial
simpl
ext
abl
eau
I
nit
ialsi
mpl
ext
abl
eau

es
abl
umn

i
on
on

ackvar

y
umns
umn

t
t

si

i
i
tperuni

onquant
escol
ut

Realordeci
corSol
ecol

col
Sl
umn

abl
abl

umn
ofi

i
i

ut
col

Basi

var
i
Pr

var

Sol
col
Cj 300 250 0 0 0 Pr
ofi
tperuni
trow
X1 X2 S1 S2 S3
SV Q
40 R1
0 S1 2 1 1 0 0

45 Const
raint R2
0 S2 1 3 0 1 0 equat
ionrows
0 S3 1 0 0 0 1 12 R3

Zj 0 0 0 0 0 0 Gr
ossPr
ofi
trow
Cj-Zj 300 250 0 0 0
NetProf
itr
ow
/I
ndi
catorr
ow/
St
ep5:
Chooset i
he“ncomi
ng”or“
ent
eri
ng”var
iabl
es
Not
e:
Theent
eri
ngv
ari
abl
eist
hev
ari
abl
ethathast
hemostposi
ti
veval
uei
ntheCj-Zjr
ow
al edasi
socal
l ndi
cat
orr
ow.Ort
heent
eri
ngv
ari
abl
eist
hev
ari
abl
ethathast
he
hi
ghestcontr
ibut
iont
oprof
itperuni
t.
a. X1inourcaseist
heent
eri
ngv ar
iabl
e
b.Thecol
umnassoci
atedwi
tht
heent
eri
ngv
ari
abl
ei edkeyorpi
scal
l votcol
umn
(X1col
umni
nourcase)

30
St
ep6
Chooset l
he“eavi
ng“ out
or“ goi
ng”var
iabl
e
==>I
nthi
sst
ep,wedet
ermi
net
hev
ari
abl
ethatwi
l
lleav
ethesol
uti orX1 (
onf or
ent
eri
ngv
ari
abl
e)
Not
e:
Ther
ow wi
tht
hemi
nimum orl
owestposi
ti
ve(
non-
negat
ive)r
epl
acementr
ati
o
showst
hev
ari
abl
etol
eav
ethesol
uti
on.

Repl
acementRat
io(
RR)=Sol
uti
onQuanti
ty(
Q)
Corr
espondi
ngval
uesi
npi
votcol
umn

Not
e:RR>0
Thevar
iabl
eleav
ingt
hesol
uti
oni edl
scal
l eavi
ngvar
iabl
eorout
goi
ngvar
iabl
e.
Ther
owassoci
atedwi
tht
hel
eav
ingv
ari
abl
eiscal
l
edkeyorpi
votr
ow(
s3col
umni
n
ourcase)
Theel
ementt
hatl
i
esatt
hei
nter
sect
ionoft
hepi
votcol
umnandpi
votr
owi
scal
l
ed
pi
votel
ement
(No1i
nourcase)

St
ep7
Repeatst
ep3-
5ti
l
lopt
imum basi
cfeasi
blesol
uti
oni
sobt
ained.
Or
:repeatst
ep3-
5ti
l
lnoposi
ti
vev
alueoccur
si heCj-Zjr
nt ow.
Not
e:
Di
vi
deeachel
ementoft
hepi
votr
owbyt
hepi
votel
ementt
ofi
ndnewv
aluesi
nthe
keyorpi
votr
ow.
Per
for
mrowoper
ati omakeal
onst lot
herent
ri
esf
ort
hepi
votcol
umnequalt
ozer
o.
2ndsi
mpl
ext
abl
eau

Cj 300 250 0 0 0

SV X1 X2 S1 S2 S3 Q

16
0 S1 0 1 1 0 -
2 R’
1=R1-
2R3

33
0 S2 0 3 0 1 -
1 R’
2=R2-
R3

30 12 R’
3=R3
X1 1 0 0 0 1
0 1
Zj 300 0 0 0 300 3600
Cj-Zj 0 250 0 0 -
300

31
 3rdsi
mpl
ext
abl
eau

Cj 300 250 0 0 0

SV X1 X2 S1 S2 S3 Q

0 S1 0 0 1 -
1/3 -
5/3 5 R’
’
1=R’
1-
R’2

25
X2 0 1 0 1/
3 -
1/3 11 R’
’
2=R2/
3
0
R’
’
3=R’
3
30
X1 1 0 0 0 1 12
0
300 250 0 250/
3
Zj 6350
650/
3
Cj-Zj 0 0 0 -
250/
3 -650/
3 Si
nceal
lthe
Cj -Zj < 0
opt
imal
sol
uti
oni
sreachedat
.
Ther
efor
e,X1=12,
X2=11,
S1=5andMaxZ=6350

Exer
cise:
AJui
ceCompanyhasav
ail
abl
etwoki
ndsoff ces:Or
oodJui angeJui
ceandGr
ape
Jui
ce.Thecompanypr
oducest
wot
ypesofpunches:PunchA andPunchB.One
bot
tl
eofpunchAr
equi
res20l
i
ter
sofOr
angeJui
ceand5l
i
ter
sofGr
apeJui
ce.
1
Bot
tl
eofpunchBr
equi
res10l
i
ter
sofOr
angeJui
ceand15l
i
ter
sofGr
apeJui
ce.

Fr
om eachofbot
tl
eofPunchAapr
ofi
tof$4i
smadeandf
rom eachbot
tl
eofPunch
Bapr
ofi
tof$3i
smade.
Supposet
hatt
hecompanyhas230l
i
ter
sofOr
angeJui
ce
and120l
i
ter
sofGr
apeJui
ceav
ail
abl
e

Requi
red:
a.For
mul
atet
hispr
obl
em asaLPP
b.Howmanybot
tl
esofPunchAandPunchBt
hecompanyshoul
dpr
oducei
nor
dert
o
maxi
mizepr
ofi
t?(
Usi
ngt
hesi
mpl
exmet
hod)
c.Whati
sthi
smaxi
mum pr
ofi
t?

Sol
uti
on:
Jui
ceneededf
oronebott
leof
Jui
ce PunchA PunchB Jui
ceAv
ail
abl
e
_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
__
Or
angeJui
ce(
lt
) 20 10 230

32
Gr
apeJui
ce(
lt
) 5 15 120
Pr
ofi
tpert
ent $4 $3
_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
_

heNoofbot
LetX1=t tl
esofpunchApr
oduced.
heNoofbot
X2=t tl
esofpunchBpr
oduced.

LPPModel
MaxZ=4X1+3X2
St
:
20X1+10X2 <230 OrangeConstraint
5X1+15X2 <120 Gr
apeConstraint
X1, X2 >0 Non-
negat
ivi
tyconst r
aint
St
andar
dfor
m
Max.
Z=4x1+3x2+0s1+0s2+0s3
St
:
20x1+3x2+s1+0s2=230
Standar
d
5x1+15x2+0s1+s2+=120 form
x1 ,x2 , s1,s2, >0
e,
Wher s1=Unusedor
angejui
ce
s2=Unusedgr
apejui
ce

I
nit
ialsi
mpl
ext
abl
eau

Cj 4 3 0 0

SV X1 X2 S1 S2 Q

0 S1 20 10 1 0 230

0 S2 5 15 0 1 120
Zj 0 0 0 0 0
Cj-Zj 4 3 0 0

33
 2ndsi
mpl
ext
abl
eau

Cj 4 3 0 0

SV X1 X2 S1 S2 Q

4 X1 1 1/
2 1/
20 0 11.
5

0 S2 0 25/
2 -
1/4 1 62.
5
Zj 4 2 1/
5 0 46
Cj-Zj 0 1 -
1/5 0

Opt
imalsi
mpl
ext
abl
eau

Cj 4 3 0 0

SV X1 X2 S1 S2 Q

4 X1 1 0 3/
50 -
1/25 9

0 X2 0 1 -
1/50 2/
25 5
4 3 0.
12
Zj 51
0.
08
0 0 -0.
12 -
Cj-Zj
0.
08 X1=9bot
tl
esofpunch
A
X2=5bott
lesofpunchB
s1=0
s2=0
MaxZ=$51

Exer
cise:
Sol
vet
hef
oll
owi
ngLPPsusi
ngt
hesi
mpl
exmet
hod
1.Max.
Z=3x1+5x2 2. Z=20x
Max. 1+10x2
Subj
ectt
o: Subj
ectt
o:
x2 <6 5x1+4x2<250
3x1+2x2<18 2x1+5x2<150
x1,x2 >0 x1,x2 >0
Answer: Answer :

34
x1=2,
x2=6,s1=0,
s2=0andMaxZ=$36 x1=50,
x2=0,
s1=0,
s2=50and
MaxZ=$1,000

3.
2.MI
NIMI
ZATI
ONPROBLEMS
 Mi
nimi
zeZwi
thi
nequal
i
tiesofconst
rai
ntsi
n“>“
for
m
Ther
ear
etwomet
hodst
osol
vemi
nimi
zat
ionLPpr
obl
ems:
1.Di
rectmet
hod/
BigM-
met
hod/
 Usi
ngar
ti
fi
ci
alvar
iabl
es
2.Conv
ersi
onmet
hod
 Mi
nimi
zat
ionbymaxi
mizi
ngt
hedual
 Sur
plusVar
iabl
e(-
s):
 Av
ari
abl
einser
tedi
nagr
eat
ert
hanorequal
toconst
rai
ntt
ocr
eat
e
equal
i
ty.I
trepr
esent
stheamountofr
esour
ceusageabov
ethemi
nimum r
equi
red
usage.
 Sur
plusv
ari
abl
eissubt
ract
edf
rom a>const
rai
nti
nthepr
ocessof
conv
ert
ingt
heconst
rai
ntt
ost
andar
dfor
m.
 Nei
thert
hesl
acknort
hesur
plusi
snegat
ivev
alue.Theymustbe
posi
ti
veorzer
o.

Exampl
e:
1. 2x1+x2<40==>i
saconst
rai
nti
nequal
i
ty
x1=12andx2=11==>2x1+x2+s=40==>2(12)+11+s=40
==>s=5unusedr esour
ce
2. 5x1+3x
2<45
x1=12andx2=11==>5x1+3x2+s=45==>5(12)+3(11)
+s=45

==>s=0unusedr
esour
ce(
Noi
dler
esour
ce)
3. 5x1+2x
2>20
x1=4.5andx2=2==>5x1+2x2-s=20==>5( 4.5)
+2(2)
-s=20
==>s=6unusedresource
4. 2x1+x2>40
x1=0andx2=0(Noproduction)==>5x1+2x2-s=20==>5(4.
5)+2(
2)s=20
-
==>s=-
6(Thisismathemati
call
yunaccept
ed)
Thus,
inor
dert
oav
oidt
hemat
hemat
ical
cont
radi
cti
on,
wehav
etoaddar
ti
fi
cial
v
ari
abl
e(A)

 Ar
ti
fi
cialvar
iabl
e(A)
:
Ar
ti
fi
ci
alv
ari
abl
eisav
ari
abl
ethathasnomeani
ngi
naphy
sical
sensebutact
sasa
t
ool
tocr
eat
eani
nit
ial
feasi
bleLPsol
uti
on.

35
Note:
Typeofconstrai
nt Toputi ntostandardfor
m
< -
--
--
--
--
--
---
--
--
--
---
--
--
--
--
--
--
--
--
--
--
--Addaslackvari
able
= -
--
--
--
--
--
---
--
--
--
---
--
--
--
--
--
--
--
--
--
--
--
Addanarti
fi
cial
variabl
e
> -
--
--
--
--
--
---
--
--
--
--Subtr
actasurplusv
ari
ableandaddar ti
fi
cial
var
iabl
e

1.
BigM-
met
hod
/
Char
nesPenal
tyMet
hod/
TheBi
g-M Met
hodi
samet
hodwhi
chi
susedi
nremov
ingar
ti
fi
cialv
ari
abl
esf
rom t
he
basi
s.I
nthi
smet
hod;weassi
gncoef
fi
cient
stoar
ti
fi
cialv
ari
abl
es,undesi
rabl
efr
om
t
heobj
ect
ivef
unct
ionpoi
ntofv
iew.I
fobj
ect
ivef
unct
ionZi
stobemi
nimi
zed,t
hena
v
eryl
arge posi
ti
ve pr
ice ( ed penal
cal
l ty)i
s assi
gned t
o each ar
ti
fi
cialv
ari
abl
e.
Si
mil
arl
y,i
fZi
stobemaxi
mized,
thenav
eryl
argenegat
ivepr
ice(
al edpenal
socal
l ty)
i
sassi
gnedt
oeachoft
hesev
ari
abl
es.

Fol
l
owi
ngar
ethechar
act
eri
sti
csofBi
g-M Met
hod:
a.Hi
ghpenal
tycost(
orpr
ofi
t)i
sassumedasM
b.M i
sassi
gnedt
oar
ti
fi
ci
alv
ari
abl
eAi
ntheobj
ect
ivef
unct
ionZ.
c.Bi
g-M met
hodcanbeappl
i
edt
omi
nimi
zat
ionaswel
lasmaxi
mizat
ionpr
obl
ems
wi
tht
hef
oll
owi
ngdi
sti
nct
ions:

i
.Mi
nimi
zat
ionpr
obl
ems
-
Assi
gn+M ascoef
fi
cientofar
ti
fi
cialv
ari
abl
eA i
nthe
obj
ect
ivef
unct
ionZ
i
i.Maxi
mizat
ionpr
obl
ems:
-
Her
e–M i
sassi
gnedascoef
fi
cientofar
ti
fi
cialv
ari
abl
eA
i
ntheobj
ect
ivef
unct
ionZ
d.Coef
fi
cientofS(
slack/
sur
plus)t
akeszer
oval
uesi
ntheobj
ect
ivef
unct
ionZ
e.Formi
nimi
zat
ion pr
obl
em,t
he i
ncomi
ng v
ari
abl
e cor
responds t
othe hi
ghest
negat
iveval
ueofCj-
Zj.
f
. Sol
uti
oni
sopt
imalwhent
her
eisnonegat
ivev
alueofCj-
Zj.
(Formi
nimi
zat
ion
case)
Exampl
e:
1.Mi
ni zeZ=25x1+30x2
mi
Subj
ectt
o:

36
 20x1+15x2>100
2x1+3x2 >15
x1, x2 >0
Solut
ion
Step1
St
andar
dizet
hepr
obl
em
Mi
ni zeZ=25x1+30x2+0s1+0s2+MA1+MA2
mi
Subj
ectt
o:
20x1+15x2-s1+A1=100
2x1+ 3x2–s2+A2 =15
x1,x2 ,
s1,s2,A1,
A2>0

St
ep2
I
nit
ialsi
mpl
ext
abl
eau
Thei
nit
ial
basi
cfeasi
blesol
uti
oni
sobt
ainedbyset
tngx1=x2=s1=s2=0
i
Nopr
oduct
i x1=x2=s1=0==>20(
on, 0)+15(
0)-0+A1=100==>A1=100
x1=x2=s2=0==>0(
0)+3(
0)-0+A2=15==>A2=15

I
nit
ialsi
mpl
ext
abl
eau

Cj 25 30 0 0 M M

SV X1 X2 S1 S2 A1 A2 Q RR

M A1 20 15 -
1 0 1 0 100 100/
20=5

M A2 2 3 0 -
1 0 1 15 15/
2=7.
5

Zj 22M 18M M
- M
- M M 115M
Cj-Zj 25-
22M 30-18M M M 0 0

Not
e:
Oncean ar
ti
fi
cialvar
iabl
e hasl
eftt
hebasi
s,i
thasser
vedi
tspur
poseandcan
t
her
efor
e be r
emov
ed f
rom t
he si
mpl
ex t
abl
eau.An ar
ti
fi
cialv
ari
abl
eis nev
er
consi
der
edf
orr
e-ent
ryi
ntot
hebasi
s.

2ndSi
mpl
exTabl
eau

37
Cj 25 30 0 0 M

SV X1 X2 S1 S2 A2 Q

25 X1 1 3/
4 -
1/20 0 0 5 R’
1=R1/
20

M A2 0 3/
2 1/
10 -
1 1 5 R’
2=R2-2R’

Zj 25 75/
4+3/
2M -
5/ 10M -
4+1/ M M 125+5M
0 45/
4-3/
2M 5/
4- 10M
1/ M
Cj-Zj
0

3rdSi
mpl
exTabl
eau

Cj 25 30 0 0

SV X1 X2 S1 S2 Q

25 X1 1 0 -
1/10 1/
2 5/
2 R’
’
1=R’
1-
3/4R’
’
2

30 X2 0 1 1/
15 -
2/3 10/
3 R’
’
2=R’
2/
3/2

Zj 25 30 -
1/2 -
15/
2 162.
5
0 0 1/
2
Cj-Zj
15/
2

Cj-Zj>0==>Opt
imalsol
uti
oni
sreached
X1=5/2
X2=10/3andMinZ=162.5

Not
e:
Asl A”var
ongasan“ iabl
eisav
ail
abl
eint
hesol
uti
onv
ari
abl
ecol
umn,t
hesol
uti
oni
s
i
nfeasi
ble.

2.Uset
hepenal
ty(
Big-
M)met
hodt
osol
vet
hef
oll
owi
ngLPP
nZ=5x1+3x2
Mi
Subj
ectt
o:
2x1+4x2<12

38
 2x1+2x2 =10
5x1+2x2 >10
x1, x2 >0
Sol
uti
on
MinZ=5x1+3x2+0s1+0s2+MA1+MA2
Subj ectto: Ifnopr oduct i
on
2x1+4x2+s1=12==>x1=x2=0==>s1=0(Sol uti
onValueintheini
ti
alsi
mplextabl
eau)
2x1+2x2+A1=10==>x1=x2=0==>A1=15(Sol ut
ionValueint
heinit
ialsi
mplextabl
eau)
5x1+2x2–s2+A1=10==>x1=x2=s2=0==>A2=10(Sol uti
onValueintheini
ti
alsi
mplextabl
eau)
x1,x2 ,s1,s2,A1,
A2>0
I
nit
ialSi
mpl
ext
abl
eau

Cj 5 3 0 0 M M
RR
SV X1 X2 S1 S2 A1 A2 Q

6
0 S1 2 4 1 0 0 0 12

M A1 2 2 0 0 1 0 10 5

M A2 5 2 0 -
1 0 1 10 2
Zj 7M 4M 0 M M M 20M
Cj-Zj 5-
7M 3-4M 0 -M 0 0
2ndsi
mpl
ext
abl
eau

Cj 5 3 0 0 M

SV X1 X2 S1 S2 A1 Q

0 S1 0 16/
5 1 2/
5 0 8

M A1 0 6/
5 0 2/
5 1 6
5 X1 1 2/
5 0 -
1/5 0 2
10+6
Zj 5M 6/
5M +2 0 5M -
2/ 1 M
M
Cj-Zj 0 6/
- 5M +1 0 -
2/5M+1 0
3rd
Cj 5 3 0 0 M simplex
tabl
eau
SV X1 X2 S1 S2 A1 Q

3 X2 0 1 5/
16 1/
8 0 2.
5

M A1 0 0 -
3/8 1/
4 1 3
5 X1 0 0 -
1/8 -
1/4 0 1 39
Zj 5 3 -
3/8M +5/
6 M/
4-7/
8 M 12.
5+3M
Cj-Zj 0 0 8M -
3/ 5/6 -
M/4+7/
8 0
 4thSi
mpl
ext
abl
eau

Cj 5 3 0 0

SV X1 X2 S1 S2 Q

3 X2 0 1 1/
2 0 1

0 S2 0 0 -
3/2 1 12
5 X1 0 0 -
1/2 0 4
Zj 5 3 -
1 0 23
Cj-Zj 0 0 1 0

X1=4,
X2=1,
S1=0,
S2=12andMi
nZ=23
3.Uset
hepenal
ty(
Big-
M)met
hodt
osol
vet
hef
oll
owi
ngLPP
MaxZ=2x1+x2+3x3
Subj
ectt
o:
2x1+ x2 +x3 <5
2x1+3x2+4x3 =12
x1,x2, x3 >0

Sol
uti
on

I
nit
ialSi
mpl
ext
abl
eau

Cj 2 1 3 0 -M
RR
SV X1 X2 X3 S1 A1 Q

2.
5
0 S1 1 1 2 1 0 5

-
M A1 2 3 4 0 1 12
3
40
Zj -
2M -
3M -
4M 0 -
M -
12M
Cj-Zj 2M+2 3M+1 4M+3 0 0
 2ndsi
mpl
ext
abl
eau

Cj 2 1 3 0

SV X1 X2 X3 S1 A1 Q RR

5
3 X3 1/
2 1/
2 1 1 0 5
2
2 3 4 0
-
M A1 12
1
Zj 3/2 -
M +3/
2 3 2M+3/
2 -
M -
2M+15/
2
Cj-Zj 1/22 M-
1/2 0 -
2M-
3/2 0

3rdsi
mpl
ext
abl
eau
Cj 2 1 3 0
RR
SV X1 X2 X3 S1 Q

6
3 X3 1/
2 0 1 3/
2 1.
5

1 X2 0 1 0 -
2 2 Notdef
ined

Zj 3/2 1 3 5/
2 13/
2
Cj-Zj 1/2 0 0 -
5/2

4thsi
mpl
ext
abl
eau

Cj 2 1 3 0

SV X1 X2 X3 S1 Q

3 X1 1 0 2 3 3

1 X2 0 1 0 -
2 2
41
Zj 2 1 4 4 8
Cj-Zj 0 0 -
1 -
4
 Cj-Zj<0==>opt
imalsol
uti
on

X1=3,
X2=2,
X3=0,
S1=0andMaxZ=8
3.
3.SPECI
ALCASESI
NSI
MPLEXMETHOD

1.Mi
xedconst
rai
nts
Exampl
e
MaxZ=6x1+8x2
Subj
ectt
o:
x2 <4
x1+ x2 =9
6x1+2x2 >24
x1,x2 >0

St
andar
dfor
m
Max.
Z=6x1+8x2+0s1+0s2+0s3-
M A2-M A3
St
:
x2+s1 =4
x1+x2+ A2 =9
6x1+2x2 -s3 +A3 =24 St
andar
dfor
m

Al
lVar
iabl
es >0

I
nit
ialsi
mpl
ext
abl
eau

42
 Cj 6 8 0 0 -
M -
M

SV X1 X2 S1 S3 A2 A3 Q

0 1 1 0 0
0 S1 4
0
1 1 0 0 1
-
M A2 9
0
6 2 0 -
1 0
-
M A3 4
1
Zj -
7M -
3M 0 +M -
M -
M 24
Cj-Zj 7M +6 3M+8 0 -
M 0 0

he4tht
Ans:Att abl
eau:X1=5,
X2 =4 ,
S3 =14andMaxZ=62

Not
e:
Fort
hei
nit
ial
basi
s,usear
ti
fi
cialvar
iabl
esf
orconst
rai
ntst
hathav
ethem.Ot
her
wise,
useaconst
r ntsl
ai ackvar
iabl sur
e.Hence, plusvar
iabl
eswi
l
lnotappeari
nani
nit
ial
sol
uti
on.

2.Twoi
ncomingvari
ables
/OrTi
eforenter
ingvari
abl
es/
I
nor
dert
obr
eakt
hist
ie,
thesel
ect
ionf
ort
hekeycol
umn(
ent
eri
ngv
ari
abl
e)canbe
madear
bit
rar
y.Howev
er;
thenumberofsol
uti
oncanbemi
nimi
zedbyadopt
ingt
he
f
oll
owi
ngr
ules:
1.I
fthereisatiebetweentwodeci si
onvari
abl
es,thent hesel
ectioncanbemade
arbitr
ary.
2.I
fthereisatiebetweenadeci si
onvari
abl
eandasl ack(orsurplus)var
iabl
e,t
hen
selectthedeci
sionvar
iabletoenteri
ntobasisfi
rst.
3.I
fthereisatiebetweenslackorsurpl
usvariabl
e,thenselecti
oncanbemade
arbit
rary.

Exampl
e:

43
I
ftheequat
ioni
smaxZ:

Cj

SV X1 X2 S1 S3 Q

Zj
Cj-Zj 5 2 5 0

I X1i
nsuchacase, stheent
eri
ngv
ari
abl
e

3.I
nfeasi
bil
it
y
A si
tuat
ion wi
th no f
easi
ble sol
uti
on mayexi
sti
fthe pr
obl
em was f
ormul
ated
i
mpr
oper
ly.

I
nfeasi
bil
i
tycomesaboutwhent
her
eisnosol
uti
ont
hatsat
isf
iesal
loft
hepr
obl
em’
s
const
rai
nts.

I
nthesi
mpl
exmet
hod,ani
nfeasi
blesol
uti
oni
sindi
cat
edbyl
ooki
ngatt
hef
inal
t
abl
eau.
Ini
t lCj-Zjr
,al owent
ri
eswi
l
lbet
hepr
opersi
gnt
oi yopt
mpl imal
it
y,butan
ar
ti
fi
cialvar
iabl
e(A)wi
l
lst
il
lbei
nthesol
uti
onmi
x.

Exampl
e:
Mi
nimi
zat
ioncase

44
 Cj 5 8 0 0 M

SV X1 X2 S1 S2 A2 Q

5 X1 1 1 -
2 3 0 200

8 X2 0 1 1 2 0 100
M A2 0 0 0 -
1 1 20
Zj 5 8 -
2 31-
M M 1,
800+200M
Cj-Zj 0 0 2 M-
31 0

Ev
ent lCj-Zjar
houghal eposi
ti
veor0(
i.
ethecr
it
eri
onf
oranopt
imalsol
uti
oni
na
mi
nimi
zat
ioncase)
,nof
easi
blesol
uti
oni
spossi
blebecauseanar
ti
fi
cialvar
iabl
e(A2)
r
emai
nsi
nthesol
uti
onmi
x.

4.UnboundedSol
uti
ons
Nof
ini
tesol
uti
onmayexi
sti
npr
obl
emst
hatar
enotbounded.
Thi
smeanst
hata
v
ari
abl
ecanbei
nfi
nit
elyl
argewi
thoutv
iol
ati
ngaconst
rai
nt.

I
nthesi
mpl
exmet
hod,t
hecondi
ti
onofunboundednesswi
l
lbedi
scov
eredpr
iort
o
r
eachi
ngt
hef
inalt
abl
eau.Wewi
l
lnot
ethepr
obl
em whent
ryi
ngt
odeci
dewhi
ch
v
ari
abl
etor
emov
efr
om t
hesol
uti
onmi
x.

Thepr
ocedur
einunboundedsol
uti
oni
stodi
vi
deeachquant
it
ycol
umnnumberbyt
he
cor
respondi
ngpi
votcol
umnnumber
.Ther
ow wi
tht estposi
hesmal
l ti
ve r
ati
ois
r
epl
aced.Buti
ftheent
ir
erat
iost
ur obenegat
noutt iveorundef
ined,i
tindi
cat
est
hat
t
hepr
obl
em i
sunbounded.

Cj 6 9 0 0
Exampl
e:
SV X1 X2 S1 S2 Q Maxi
mizat
ion
case
9 X2 -
1 1 2 0 30

0 S2 -
2 0 -
1 1 10

Zj -
9 9 18 0 270 45

Cj-Zj 15 0 -
18 0
Thesol
uti
oni
ntheabov
ecasei
snotopt
i lCj-Zjent
malbecausenotal ri e0or
esar
negat
ive,asr
equi
redi
namaxi
mizat
ionpr
obl
em.Thenextv
ari
abl
etoent
ert
he
sol
uti dbeX1.
onshoul Todet
ermi
newhi
chv
ari
abl
ewi
l
lleav
ethesol
uti
on,
weexami
ne
t
her
ati
osoft
hequant
it
ycol
umnnumber
stot
hei
rcor
respondi
ngnumber
si heX1or
nt
pi
votcol
umn.Si
ncebot
hpi
votcol
umnnumber
sar
enegat
ive,anunboundedsol
uti
on
i
sindi
cat
ed.

Nounboundedsol
uti
ons,
noout
goi
ngv
ari
abl
ewi
l
lexi
st.

5.Degener
acy
Ti
/ ef
orl
eavi
ngbasi
cvar
iabl
e(keyr
ow)
/

I
fther
eisat
ief
ort
he smal
l
estr
ati
o,t
hisi
s a si
gnalt
atdegener
acyexi
sts.
Degener
acycanoccurr
ighti
nthef
ir
st(
ini
ti
alt
abl
eau)
.Thi
snor
mal
l
yhappenswhen
t
henumberofconst
rai
ntsi
slesst
hant
henumberofv
ari
abl
esi
ntheobj
ect
ive
f
unct
ion.Pr
obl
em canbeov
ercomebyt
ri
alander
rormet
hod.

Cj 5 8 2 0 0 0

Q
SV X1 X2 X3 S1 S2 S3 RR

8 X2 1/
4 1 1 -
2 0 0 10 10/
1/4=40

20/
4=5 Ti
efort
hesmal
lestr
ati
0 S2 4 0 1/
3 -
1 1 0 20 i
ndi
cat
esdegener
acy.
2 0 2 2/
5 0 10/
2=5
0 S3 10
1
2 8 8 16 0
Zj 80 46
0
Cj-Zj 3 0 -
6 -
16 0
0
Degener
acycoul
dleadt
oasi
tuat
ionknownascy
cli
ng,
inwhi
cht
hesi
mpl
exal
gor
it
hm
al
ter
nat
ivesbackandf
ort
hbet
weent
hesamenon-
opt
imalsol
uti
ons,
i.
e,i
tput
sanew
v
ari
abl
ein,
thent
akesi
touti
nthenextt
abl
eau,
put
sitbacki
n,andsoon.
Onesi
mpl
ewayofdeal
i
ngwi
tht
hei
ssuei
stosel
ectei
therr S2 orS3i
ow( nthi
scase)
ar
bit
rar
y.I
fwear
eunl
uckyandcy
cli
ngdoesoccur
,wesi
mpl
ygobackandsel
ectt
he
ot
herr
ow.
Remar
k
Whent
her
eisat
iebet
weenasl
ackandar
ti
fi
cialv
ari
abl
etol
eav
ethebasi
s,t
he
pr
efer
enceshal
lbegi
vent
oar
ti
fi
cialv
ari
abl
etol
eav
ethebasi
sandt
her
eisnoneed
t
oappl
ythepr
ocedur
eforr
esol
vi
ngsuchcases.

6.Mul
ti
pleOpt
imalSol
uti
ons
Mul
ti
pleopt
imalsol
uti
onsexi
stwhennon-
basi
cvar
iabl
econt
ainszer
ooni
ts
Cj-Zjr
ow.

Exampl
e:
Max
imi
zat
ionpr
obl
em

Cj 3 2 0 0

SV X1 X2 S1 S2 Q

2 X2 3/
2 1 1 0 6

0 S2 1 0 1/
2 1 3

Zj 3 2 2 0 12

Cj-Zj 0 0 -
2 0

MaxZ=3X1+2X2

X1=0,
X2=6,
S2=3andMaxZ=12or
:X1=3,
X2=3/
2andMaxZ=12

47
TheCj-Zjv
alueoft
heNon-
basi
cvar
iabl
e(X1)i
s0.
Thus,t
her
eisal
ter
nat
iveopt
imal
sol
uti
on.

SUMMARYOFLP:SI
MPLEXMETHOD

1.The st
andar
d f
orm of LP pr
obl
em shoul
d hav
e t
he f
oll
owi
ng
char
act
eri
sti
cs:
I
.Al
ltheconst
rai
ntsshoul
dbeexpr
essedasequat
ionsbysl
ackorsur
plus
and/
orar
ti
fi
cial
var
iabl
es
I
I.Ther
ighthandsi
deofeachconst
rai
ntshoul
dbemadenon-
negat
ive;i
f
i
tisnot
,thi
sshoul
dbedonebymul
ti
ply
ingbot
hsi
desoft
her
esul
ti
ng
const
rai
ntby-
1.
Exampl
e:
2X1+3X2-
4X3+X3<-
50,
wemul
ti
plybot
hsi
desbynegat
ive
I
II
.Thr
eet
ypesofaddi
ti
onal
var
iabl
es,
namel
y
a.Sl
ackVar
iabl
e(S)
b.Sur
plusvar
iabl
e(-
S),
and
c.Ar
ti
fi
cialvar
iabl
es(
A)
ar
eaddedi
nthegi
venLPpr
obl
em t
oconv
erti
tint
ost
andar
dfor
mfor
t
wor
easons:
i
.toconv
ertani
nequal
i
tyt
ohav
east
andar
dfor
m ofanLP
model
,and

48
 i
i
.togetani
nit
ialf
easi
blesol
uti
onr
epr
esent
edbyt
hecol
umns
ofani
dent
it
ymat
ri
x.

Thesummer
yoft
heext
rav
ari
abl
esneededt
oaddi
nthegi
venLPpr
obl
em t
oconv
ert
i
tint
ost
andar
dfor
misgi
venbel
ow:

Tabl
e
Coeff
ici
entofextr
avari
ables Presenceofvari
abl
esin
Typesof
Ext
ravar
iabl
est
obeadded i
ntheobject
ivefunct
ion theini
ti
alsol
uti
onmix
Const
rai
nt
MaxZ Mi
nZ
< Addonl
ysl
ackvar
iabl
e 0 0 Yes
> Subt
ractsur
plusvar
iabl
eand 0 0 No
Addar
ti
fi
cialvar
iabl
e -
M +M Yes
= Addar
ti
fi
cialvar
iabl
e -
M +M Yes

2.Testofopt
imal
it
y
i
.Ifal
lCj-Zj<0,
thent
hebasi
cfeasi
blesol
uti
oni
sopt
imal
(Maxi
mizat
ioncase)
i
i
.Ifal
lCj-Zj>0,
thent
hebasi
cfeasi
blesol
uti
oni
sopt
imal
(Mi
nimi
zat
ioncase)
3.Var
iabl
etoent
ert
hebasi
s
i
.Av
ari
abl
ethathast
hemostposi
ti
vev
aluei
ntheCj-Zjr
ow
(
Maxi
mizat
ion case)
i
i
.Av
ari
abl
ethathast
hehi
ghestnegat
ivev
aluei
ntheCj-Zjr
ow(
Mini
mizat
ion
case)
4.Var
iabl
etol
eavet
hebasi
s
Ther
owwi
tht
henon-
negat
iveandmi
nimum r
epl
acementr
ati
o
(
Forbot
hmaxi
mizat
ionandmi
nimi
zat
ioncases
i
. RR>0
e:

49
 CHAPTERI
V
SENSATI
VITYANALYSI
SANDDUALI
TY

4.
1.SENSATI
VITYANALYSI
S
Sensi
ti
vi
tyAnal
ysi
sisconcer
nedwi
tht
hest
udyof‘
Sensi
ti
vi
ty‘
oft
heopt
imalsol
uti
on
ofanLPPwi
thdi
scr
eti
onv
ari
abl
es(
changes)i
npar
amet
ers.
Thedegr
eeofsensi
ti
vi
ty
oft
he sol
uti
on due t
othose v
ari
ati
ons can r
ange f
rom no change atal
ltoa
subst
ant
ialchangei
ntheopt
imalsol
uti
on oft
hegi
ven LPP.Thus,i
nsensi
ti
vi
ty
anal
ysi
s,wedet
ermi
net
her
angeov
erwhi
cht
heLPmodelpar
amet
erscanchange
wi
thoutaf
fect
ingt
hecur
rentopt
imal
sol
uti
on.Thepr
ocessofst
udy
ingt
hesensi
ti
vi
ty
oft
heopt
imal
sol
uti
onofanLPPi edpost
scal
l -opt
imalanal
ysi
s.

Thet
wosensi
ti
vi
tyanal
ysi
sappr
oachesar
e:
I
.Tr
ial
ander
rorappr
oachand
I
I.Anal
yti
cal
appr
oach

4.
1.1.Anal
yti
calappr
oach
Fi
vet
ypesofdi
scr
etechangesi
ntheor
igi
nalLPmodelmaybei
nvest
igat
eddur
ing
t
hesensi
ti
vi
tyanal
ysi
s:
A.Changesoft
hecoef
fi
cient
soft
heobj
ect
ivef
unct
ion(
cj)
B.Changesoft
heRHSQuant
it
y(bj)
C.Changesoft
hei
nput
-out
putcoef
fi
cient

50
 D.Add/
del
eteconst
rai
nts
E. Theaddi
ti
onofanewvar
iabl
etot
hepr
obl
em

A.Changesoft
hecoef
fi
cient
soft
heobj
ect
ivef
unct
ion(
cj)
Deci
sionv
ari
abl
escanbe:
i
.Basic(int
hesoluti
on)
i
i.Non-basi
c(out
-ofthesol
uti
on)

Not
e:
I
nst
eadofr
esol
vi
ngt
heent
ir
epr
obl
em asanew pr
obl
em wi
thnew par
amet
ers,
wemayt
aket
heor
igi
nalopt
imalsol
uti
ont
abl
easani
nit
ialsol
uti
ont
abl
efort
he
pur
poseofknowi
ngr
angesbot
hlowerandupperwi
thi
nwhi
chapar
amet
ermay
assumev
alue.

a.Rangef
ort
hecoef
fi
cient
sofbasi
cdeci
sionvar
iabl
es
The r
ange ofopt
imal
it
yist
he r
ange ov
erwhi
ch a basi
c deci
sion v
ari
abl
e
coef
fi
cienti
ntheobj
ect
ivef
unct
ioncanchangewi
thoutchangi
ngt
heopt
imal
sol
uti
onmi
x.Howev
er,t
hischangewi
l
lchangeonl
ytheopt
imalv
alueoft
he
obj
ect
ivef
unct
ion.
Exampl
e:
Z=5x1+4.
Max. 5x2+x3
Subj
ectt
o:
15x1+15.
8x2 <150
5x1+6.
4x2+15x3<77
2.8x2+11.8x3<36
x1,x2 ,x3>0

Theopt
imal
tabl
eauf
ort
hissol
uti
oni
s:

Cj 5 4.
5 1 0 0 0

SV X1 X2 X3 S1 S2 S3 Q

5 X1 1 1.
053 0 0.
067 0 0 10

1 X3 0 0.
67 1 -
0.022 0.
067 0 1.
8

0 S3 0 1.
924 0 0.
258 -
0.773 1 15.
12
Zj 5 5.
342 1 0.
311 0.
067 0 51.
8

51
 Cj-Zj 0 -
0.842 0 -
0.311 -
0.067 0

X2 S1 S2 Non-
basi
c
Cj-Zjrow -
0.842 -
0.311 -
0.067
X1 row 1.
053 0.
067 0
Cj-Zjr
ow
X1row -
0.8 -
4.64 -
∞

Det
ermi
ne t
he r
ange ofopt
imal
i
tyf
ort
he coef
fi
cient
s oft
he basi
c-deci
sion
v
ari
abl
es.

Sol
uti
on:
Anal
ysi
sofbasi
cdeci
sionv
ari
abl
es
Theanal
ysi
swi
l
lbeconduct
edonpr sonX1 andX3whi
oduct char
eint
hebasi
c
sol
uti
on.Di
vi -Zjr
deeachCj owent
ryf
orv
ari
abl
esnoti
nthesol
uti
on(
fori
nst
ance,
byX2,S1andS2v
alues) byt
heassoci
atedv
ari
abl
eaijf
rom X1orX3r
ow.

I
.Anal
ysi
sofX1

St
eps:
a. TaketheCj-Zjr
owoft
heopti
malsol
uti
onoft
henon-
basi
cvar
iabl
es
b. TaketheX1rowoft
henon-
basi
cvari
abl
es
c.Cj-Zjrow
X1r
ow
 UpperLi
mit
Thesmal
lestposi
ti
venumberi
ntheCj-Zjr
owt
ell
showmucht
hepr
ofi
t
X1
ofX1canbei
ncr
easedbef
oret
hesol
uti
oni
schanged.

UpperLi
mit=Cj(forX1)+t
hesmall
estposi
ti
veval
ueofCj-Zjr
ow
=5+∞=∞ X1r
ow
Note:Cj(
forX1)=5(Lookint
he OFoftheLPproblem)

 LowerLi
mit
Thel
argestnegat
ivenumbercl
osest(
negat
iveamountcl
osestt
o0)

LowerLi
mit
=Cj(
forX1)
+Thel
argestnegat
iveval
ueofCj-Zjr
ow

52
 =5+(
-0.
8)=4.
2 X1r
ow

Ther
efor
e,t
her
angeofopt
imal
i
tyf
ort
hecoef
fi
cientofX1 i 2<Cj(
s4. forX1)<∞
(
Thecoef f
ici
entofX1 i
ntheobjecti
vef
unctioncanchangebetween4.2and∞
wi
thoutchangingt
heoptimalsol
uti
onmixX1=10,X3=1.
8andS3=15.
12)

I
I.Anal
ysi
sofX2

 UpperLimi t
=Cj(forX3)+Thesmal
lestposi
ti
veval
ueofCj-Zjrow
=1+(14.13)=15.13 X1row
Not
e:Cj(
forX3)=5(
Lookinthe OFoftheLPprobl
em)

 LowerLi
mit
=Cj(
forX3)+Thel
argestnegat
iveval
ueofCj-Zjrow
=1+( -
1)=0 X1row

Ther
efor
e,t
her
angeofopt
imal
i
tyf
ort
hecoef
fi
cientofX3i
s0<Cj(
forX3)<
15.
13(
Thecoef
fi
cientofX3i
ntheobj
ect
ivef
unct
ioncanchangebet
ween0and15.
13
wit
houtchangingtheopti
malsoluti
onmi xX1=10,X3=1.
8andS3=15.12).
However,thi
s
changewi
llchangeonl
ytheopt
imalvalueoft
heobject
ivef
unct
ion(
i.
e.MaxZwil
lchange)

Exer
cise:
Z=50x1+120x2
Max.
Subj
ectt
o:
2x1+4x2<80
3x1+x2<60
x1,x2 >0

Determinet
her
angeofopt i
mal
i
tyf
ort hecoef
fi
cientofthebasi
cv ariabl
es.
X2 S1 S2
Opti
mal Sol
ut
Noi
o
nn-
basi
c
Cj-Zjrow -
0.842 -
0.311 -0.067
X3row 0. 67 -
0.022 0.067
Cj-Zjr
ow
X1row -1.26 14.13 -1

Cj

SV X1 X2 S1 S2 Q

5 X1 1/
2 1 1/
4 0 20

53
 1 S2 5/
2 0 -
1/4 1 40
Zj 60 120 30 0 $2,
400
Cj-Zj $-
10 $0 $-
30 $0

Ans:Ther
angeofopt
imal
it
yforX2’
spr
ofi
tcoef
fi
cienti
s:$100<Cj(
forX2)<∞

b.Ther
angef
ort
henon-
basi
cvar
iabl
es
I
fther
eisav
ari eCj,notpar
abl ti
cipat
ingi
ntheopt
imalbasi
s,t
hen,i
nor
derf
ort
he
v
ari
abl
etobei
ncl nt
udedi heopt
imalsol
uti
on,i
tscoef
fi
cienti
ntheobj
ect
ivef
unct
ion
wi
l
lhav
etochangef
rom t
heexi
stngCjt
i oanewl
ev Cj(
el new)
.

Cj(
new)
>Zj

Ther
angeofi
nsi
gni
fi
cancei
sther
angeov chCjr
erwhi atesf
ornon-
basi
cvar
iabl
es
canv
arywi
thoutcausi
ngachangei
ntheopt
imalsol
uti
onmi
x(v
ari
abl
e)i
scal
l
edt
he
r
angeofi
nsi
gni
fi
cance.

Exampl
e:
Z=5x1+4.
Max. 5x2+x3
Subj
ectt
o:
15x1+15.
8x2 <150
5x1+6.
4x2+15x3<77
2.8x2+11.8x3<36
x1,x2 ,x3>0

Theopt
imal
tabl
eauf
ort
hissol
uti
oni
s:
Cj 5 4.
5 1 0 0 0

SV X1 X2 X3 S1 S2 S3 Q

5 X1 1 1.
053 0 0.
067 0 0 10

1 X3 0 0.
67 1 -
0.022 0.
067 0 1.
8

0 S3 0 1.
924 0 0.
258 -
0.773 1 15.
12
Zj 5 5.
342 1 0.
311 0.
067 0 51.
8
Cj-Zj 0 -
0.842 0 -
0.311 -
0.067 0

54
Cal
cul
atet
her
angeofi
nsi
gni
fi
cancef
ort
hecoef
fi
cientofnon-
basi
cvar
iabl
e(X2)

Sol
uti
on

Cj(forX2)=4.
5andZj(forX2)=5.342
Cj(new forX2)>5.
342==>Cj(new forX2).
Ift
hepr
ofi
tcont
ri
but
ionofX2 i
sgr
eat
ert
han
5.342,thenX2wi l
lbei
ncludedinthesoluti
on.

∞<Cj(
Thus, newf
orX2)
<5.
342i
sther
angeofi
nsi
gni
fi orX2.
cancef
Cj(
new forX2)canvarywit
hinthegivenrangewithoutcausi
ngachangei
nthe
opt
imal
solut
ionmi X1=10,
x( X1=0,
X3=1.
8,S1=S2=0andS3=15.
12).

B.Changei
ntheRightHand—Si
deQuant
it
y(RHS)
Or
Changei
ntheavai
labi
li
tyofr
esour
ce(
Capaci
ty)(
bj)

 Shadowpr
ices:
==>How muchshoul
daf
ir
m bewi
l
li
ngt
opayt
omakeaddi
ti
onalr
esour
ces
av
ail
abl
e?
Shadowpricessi
gnif
ythechangei
ntheopt
imalval
ueoftheobj
ecti
vefunct
ionfor1
uni
tincr
easesintheval
ueoftheRHSoftheconst
rai
ntt
hatrepr
esentt
heavail
abil
it
y
ofscar
cer
esour
ces.

Thenegat
iv henumberofCj-Zjr
eoft ni
ow i t
ssl
ackvar
iabl
ecol
umnspr
ovi
deas
wi
thshadow pr
ices.Or
:shadow pr
icesar
efoundi
ntheZjr
ow oft
hef
inalsi
mpl
ex
t
abl
eaui
nthesl
ackv
ari
abl
ecol
umns.

 RHSr
angi
ngi
sther
angeov
erwhi
chshadowpr
icesr
emai
nval
i
d.
Exampl
e:
Z=3x1+4x2
Max.
Subj
ectt
o:
3x1+5x2<15
2x1 + x2 <8
x2 <2
x1, x2 >0

55
 Theopt
imal
tabl
eaui
sgi
venas:
Cj 3 4 0 0 0

SV X1 X2 S1 S2 S3 Q

3 X1 1 0 -
0.143 0.
714 0 3.
57

0 S3 0 0 -
0.286 0.
428 1 1.
143

4 X2 0 1 0.
286 -
0.428 0 0.
857

Cj-Zj 0 0 -
0.714 -
0.428 0

Requi
red:
1.Det
ermi
net
heshadowpr
icef
oreachconst
rai
nt
2.Det
ermi
net
heRHSr
angesov
erwhi
cht
heshadowpr
icesar
eval
i
d

 Anal
ysi he1stconst
soft rai
nt(
S1)
Q S1 Q/S1

3.
57 -
0.143 -
24.
96

1.
143 -
0.286 -
3.99

0.
857 0.
286 3.
00

LowerLi
mit
=bj-
thesmallestposi
ti
venumberintheQ/Sjcol
umn
[
UpperLi
mit
=bj-
thel
argestnegati
venumberintheQ/Sjcol
umn
LowerLi
mit=b1-
thesmal lestposi
ti
venumberintheQ/S1column
=15-3=12
UpperLi
mit
=b1-thelargestnegati
venumberintheQ/S1column
=15-(-
3.99)=18. 99
Ther
ef e,12<b1<18.
or 99( Therangeofresource1overwhi cht
heshadow pr
ice
$0.
714peruniti
svalid).

 Anal
ysi he2ndconst
soft rai
nt(
S2)
Q S2 Q/S2

3.
57 0.
714 5

1.
143 0.
428 2.
67

56
0.
857 -
0.428 -
2

LowerLi
mit
=b2-
thesmal
lestposi
ti
venumberi
ntheQ/S2col
umn
=8-
(2.
67)
=5.
33

UpperLi
mit
=b2-
thel
argestnegat
ivenumberi
ntheQ/S2col
umn
=8-
(-
2)=10

Ther
ef e,
or 5.
33<b1<10(
Ther
angeofr
esour
ce2overwhi
cht
heshadowpr
ice$0.
428
peruni
tisval
i
d).

 Anal
ysi he3rdconst
soft rai
nt(
S3)

Q S3 Q/S3

3.
57 0 -

1.
143 1 1.
143

0.
857 0 -
Lower
Li
mit=b3-
the
smal
lestposi
ti
venumberint
heQ/S3col
umn
=2-
(1.
143)
=0.857

UpperLi mit=b3-
thel
argestnegati
venumberi nt
heQ/S3column
=2-
(-
∞)=∞
Therefore,0.857<b3<∞ ( Ther
angeofr
esour
ce3overwhi
cht
heshadowpr
ice$0per
uni
ti d)
sval
i .

Exer
cise:
Z=50x1+40x2
1.Max.
Subj
ectt
o:
3x1+5x2<150(
Assembl
yti
me)
x2 <20 (
Por
tabledi
splay)
8x1+5x2 <300(Warehousespace)
x1, x2 >0

57
 Thef
inalsi
mpl
ext
abl
eaui
s:
Cj 50 40 0 0 0

SV X1 X2 S1 S2 S3 Q

40 X2 0 1 8/
25 0 -
3/25 12

0 S2 0 0 -
8/25 1 3/
25 8

50 X1 1 0 -
5/25 0 5/
25 30
Zj 50 40 14/
5 0 26/
5 1980
Cj-Zj 0 0 -
14/
5 0 -
26/
5

Det
ermi
net
heshadowpr
icesf
ort
het
hreeconst
rai
ntsf
ort
heHi
ghTechCo.

Answer
:
TheZjv
aluesf
ort
het
hreesl
ackv
ari
abl
esar
e14/
5,0and26/
5,r
espect
ivel
y.
Thus,t
heshadow pr
icef
ort yt
heassembl i
meconst
rai
nti
s14/
5(i
.e.
1addi
ti
onal
assembl
yti
meovert
heexi
sti
ng150is$2.8)
Theshadowpri
cef
orthepor
tabledi
spl
ayconstrai
ntis0.
Theshadowpri
cef
ortheWarehousespaceconstr
ainti
s26.
5
Ther
efor
e,weseet
hatobt
aini
ngmor
ewar
ehousespacewi
l
lhav
ethebi
ggestposi
ti
ve
i
mpactonHi
ghTech’
spr
ofi
t.

Not
e:
Wi
thagr
eat
ert
hanorequalt
oconst
rai
nt,t
hev
alueoft
heshadowpr
icewi
l
lbel
ess
t
hanorequalt
ozer
obecausea1uni
tincr
easei
nthev
alueoft
heRHScannotbe
hel
pful
;itmakes i
tmor
e di
ff
icul
tto sat
isf
ythe const
rai
nt.As a r
esul
t,f
ora
maxi
mizat
ionpr
obl
em t
heopt
imalv
alueoft
heobj
ect
ivef
unct
ioncanbeexpect
edt
o
decr
easewhent
heRHSsi
deofagr
eat
ert
hanorequal
toconst
rai
nti
sincr
eased.

2.Sol
vet
hef
oll
owi
ngLPPusi
ngsi
mpl
exmet
hod.Af
ir
mthatmanuf
act
uresbot
hlawn
mowersandsnowbl ower s:
X1=thenumberoflawnmower s
X2 =t
henumberofsnowbl ower
s
Max.Z=$30x1+$80x2
Subj
ectt
o:
2x1+4x2<1000Laborhour
savai
labl
e
6x1+2x2 <1,200lbofst
eelavai
labl
e
x2 <20 snowblowerengineavai
labl
e
x1, x2 >0

58
a.Whati
st hebestproductmix?Whati
stheopt
imal
prof
it
?
Answer:
x1=100,x2=200andprofi
t=$19,
000
b.Whatar
etheshadowpr
ices?Whent
heopt
imalsol
uti
onhasbeenr
eached,whi
ch
r
esourcehasthehighestmargi
nalvalue?
Answer:
Theshadowpri
cefor1addi
ti
onall
abor=$15
Theshadowpri
cefor1addi
ti
onalpoundofsteel=0
Theshadowpri
cefor1addi
ti
onalsnowblowersenginemadeavai
labl
e=$20
Thus,snowblowerenginehavethehighestmar ginal
val
ueattheoptimalsol
uti
on.
c.Overwhatrangeineachoft heRHSv al
uesar etheseshadowsv ali
d?
Answer:
Theshadowpr i
ceforlaborhour
sisval
idfrom 800hoursto1,066.
66hours
Theshadow priceforpoundsofsteelisvali
df r
om 1,
000poundsupt oaninf
ini
te
numberofpounds
Theshadow priceforsnow blowerenginesr angesfrom 180enginesupto250
engi
nes
d.Whatar ether angesov erwhi chtheobjectiv
ef unct
ioncoeffici
ent
scanv aryfor
eachoft hetwodecisi
onvari
ables?
Answer:
Wi t
houtchangi
ngthecur
rentsol
uti
onmi x,t
heprof
itcoef
fi
cientf
orthemowerscan
rangefrom $0to40,whi
lethecoeffi
cientfort
heblowerscanr angef
rom $60to
i
nfi
nity.

4.
3.DUALI
TY
Ev
eryLPPhasanot
herLPPassoci
atedwi
thi
t,whi
chi
scal
l
edi
tsdual
.Thef
ir
st
wayofst
art
ingal
i
nearpr
obl
em i
scal
l hepr
edt imaloft
hepr
obl
em.Thesecond
wayofst
art
ingt
hesamepr
obl
em i
scal
l hedual
edt .Theopt
imalsol
uti
onsf
ort
he
pr
imaland t
he dualar
e equi
val
ent
,butt
heyar
e der
ived t
hrough al
ter
nat
ive
pr
ocedur
es.

Not
e:
 Thedualcontai
nseconomicinf
ormati
onusefult omanagement,anditmay
al
sobeeasiertosol
ve,i
nter
msoflesscomput at
ion,
thanthepr
imalprobl
em.
 Corr
espondi
ngtoeveryLPP,t
her
eisanotherLPP.
 Thegi
venpr
obl
em i
scal
l
edt
hepr
imal
.
 Ther
elat
edpr
obl
em t
othegi
venpr
obl
em i
sknownast
hedual
.
 Thedualofaduali
sthepr
imal
 I
fthepr
imalhasopt
imalsol
uti
on,
thedualwi
l
lhaveopt
imalsol
uti
on

59
  I
fthepr
imalhasnoopt
imalsol
uti
on,
thedualwi
l
lnothaveopt
imalsol
uti
on.
 Whetherwef
oll
owt
hedualorpr
imalsyst
em,t
heopt
imalsol
uti
onwi
l
lremai
n
equal
.

Tabl
e
Pr
imal
-DualRel
ati
onshi
p
Pri
mal Dual
Obj
ect
iveismini
mizat
ion Obj
ect
ivei
smaxi
mizat
ionandvi
cever
sa
>t
ypeconst
rai
nts <t
ypeconst
rai
nts

Noofcol
umns Noofr
ows
Noofr
ows Noofcol
umns
No ofdeci
sionvar
iabl
es Noofconst
rai
nts
Noofconst
rai
nts No ofdeci
sionvar
iabl
es
Coef
fi
cientofObj
ectfunct
ion RHSvalue
RHSval
ue Coef
fi
cientofObj
ectf
unct
ion

Dual
it
yAdvant
age
1.Thedual
for
m pr
ovi
desanal
ter
nat
ivef
orm
2.Thedualr
educesthecomput
ati
onaldi
ff
icul
tiesassoci
atedwithsomefor
mulati
on
3.Thedualprov
idesanimport
anteconomicinter
pretat
ionconcerni
ngt
hev al
ue of
scar
sresour
cesused.
Exampl
e:
Wr
it
ethedual
stot
hef
oll
owi
ngpr
obl
ems
Z=5x1+6x2
a.Max.
Subj
ectt
o:
2x1+3x2<3000(
Laborconst
rai
nt)
5x1+7x2 <1000(Machi
neconstr
aint
)
x1+x2 <5000 (
Marketconst
rai
nt)
x1,x2 >0
Sol
uti
on
Representpr i
malint
heconv
enti
onal
tabl
easf ol
l
ows
Dualvariabl
es x1 x2 Const
rai
nts
u1 2 3 <3000
u2 5 7 <1000

60
 u3 1 1 <500
MaxZ 5 6

Byr
efer
ri
ngt
heabov
etabl
e,dual
fort
hiscanbest
atedas:

Mi =3000u1+1000u2+500u3
nZ*
St
::
2u1+5u2+u3>5
3u1+7u2+u3>6
u1, u2,u3>0

Not
e:
1.Formaximi zi
ng, al
lconstr
aint
smustbebr oughtt
o“<”for
m
2.Forminimizing,allconst
rai
ntsmustbebroughtto“>”f
orm
3.Ift
heyarenot ,
usemul ti
pli
cat
ionf
actor-
1
4.“=”i
sani nt
er secti
onof“>”and“<“
Z=60x1+50x2
b.Max.
Subj
ectt
o:
2x1+4x2<80
3x1+2x2 <60
x1 <15
2x2 <36
x1,x2 >0
Sol
uti
on
Pr
imal
isr
epr
esent
edi
nthet
abl
easf
oll
ows:

Dualvariabl
es x1 x2 Const
rai
nts
u1 2 4 <80
u2 3 2 <60
u3 1 0 <15
u4 0 2 <36
MaxZ 60 50

Thedualf
ormis:
Mi =80u1+60u2+15u3+36u4
nZ*
St
::
2u1+3u2+u3 >60
4u1+2u2+ u4 >50
u1, u2,u3, u4 >0

61
c.Obt
aint
hedual
probl
em oft
hef
oll
owi
ngpr
imal
LPP
Mi
n.Z=x1+2x2
Subj
ectt
o:
2x1+4x2<160
x1-x2 =30
x1 >10
x1,x2 >0

Sol
uti
on
Rul
e:
1.Fora maxi mizati
on pr i
malwi th al
l< t ype const r
aint
st here exist
sa
minimizati
ondualpr oblem withall>typeconst r
aint
sandvi cever sa.Thus,
theinequali
tysignisreversedinallt
heconst r
aint
sexceptt henon-negativi
ty
conditi
ons.
2.Tr ansformt he<t ypeconst rai
nttoa>t ypeconst rai
ntbymul t
iplyi
ngt he
constrai
ntby-1.
3. Wr i
tethe=typeconst r
aintequival
enttotwoconst rai
ntsofthetype>and<.
Thest
andar
dpr
imal
LPPsoobt
ainedi
s:
Mi
n.Z=x1+2x2
Subj
ectt
o:
2x1-
- 4x2>-
160
x1-x2 >30
-x1+x2 >-30
x1 >10
x1,x2 >0
Letu1, u2,u3,andu4bet
hedual
var
iabl
escor
respondi
ngt
othef
our
const
rai
ntsi
ngi
venor
der
,thent
hedual
oft
hegi
venpr
imal
probl
em canbe
f
ormul
atedasf
oll
ows:
Max
Z* 160u1+30u2-
=- 30u3+10
St:
-
2u1+u2-u3+u4<1
-
4u1-u2+u3 <2
u1,
u2,u3,u4 >0
Letu=u2-u3,
thentheabov
edual
probl
em r
educest
othef
orm:

62
 Max
Z* 160u1+30u2-
=- 30u3+10
St
:
-
2u1+u+u4<1
-
4u1-
u <2
u1,u4 >0,uunr
est
ri
ctedi
nsi
gn
Her
eitmaybe not
ed t
hatt
he second const
rai
nti
nthe pr
imali
s equal
i
ty.
Ther
efor
e,t
hecor
respondi u2shoul
ngdual dbeunr
est
ri
ctedi
nsi
gn.
d.Obt
aint
hedual
probl
em oft
hef
oll
owi
ngpr
imal
LPP
Z=x1-
Max. 2x2+3x3
Subj
ectt
o:
2x1+x2+3x3=2
-
2x1+3x2+4x3=1
x1,x2,x3 >0
Sol
uti
on:
Si
ncet
hegi
venconst
rai
ntsar
ebot
hequal
i
ty,
bot
hthecor
respondi
ngdual
var
iabl
es,
u1andu2willbeunrest
ri
ctedi
nsi
gn.Thedual
oft
hegi
venpr
imal
LPPi
s:
MinZ*=2u1+u2
St:
-
2u1+2u2>1
u1+3u2 >-
1
3u1+4u2 >3
u1,u2,unr
estri
ctedi
nsi
gn
e.Af
ir
m manuf
act
urest
wopr
oduct
sAandBonmachi
neIandI
Iasshownbel
ow:
_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
__
Machi
ne Pr
oduct Avai
labl
eHour
s
A B
I 30 20 300
I
I 5 10 110
Pr
ofi
t/uni
t($) 6 8
___
____
___
___
___
_ _
_ __
___
____
_____
___
___
___
___
___
___
_____
____
____
__
Thetot
alt
imeavai
labl
eis300hr sonmachi
sand110hr neIandIIr
espect
ivel
y
.
ProductAandBcont
ri
but
e$6and$8peruni
trespect
ivel
y.

i
. For
mul
atet
heLPPmodelf
ort
hepr
imalanddet
ermi
net
heopt
imalsol
uti
on
usi
ngsi
mpl
exmet
hod.
i
i
.For
mul
atet
heLPPmodelf
ort
hedualanddet
ermi
net
heopt
imalsol
uti
onf
rom
t
hef
inal
pri
mal
simpl
ext
abl
eau

63
 Sol
uti
on
LetX1=Noofuni
tsofpr
oductApr
oduced
X2=Noofuni
tsofpr
oductBpr
oduced
Max
.Z=6X1+8X2
Subj
ectt
o:
30X1+20X2<300
5X1+10X2<110
X1, X2>0
Thef
inal
simpl
ext
abl
eaui
s:

Cj 6 8 0 0

SV X1 X2 S1 S2 Q

6 X1 1 0 -
1/20 -
1/10 4

8 X2 0 1 1/
10 3/
20 9

Zj 6 8 ½ 6/
10 96
Cj-Zj 0 0 -
1/2 -
6/10
Theopt
imal
sol
uti
oni
s:
X1=4unitsofproductApr
oduced
X2=9unitsofpr oducedandMax
oductBpr .Z=$96
i
i
.Letu1=CostofonehouronmachineI
u2=CostofonehouronmachineI
MinZ*=300u1+110u2
St:
30u1+5u2>6
20u1+5u2>8
u1, u2>0
NOTE:
Thevaluei
ntheCj-Zjr
owundercolumnsoftheslack/surpl
usvar
iableswi
thchange
i
nsigngivedi
rect
lyt
heopt
imalval
uesofthedual
/primalbasi
cvari
ables.
Ther
efore,

uj=-
(Cj-Zj)

Not
e: umns1cor
Thecol r ou1andt
espondst umns2cor
hecol r ou2
espondst
u1=-
$(-
1/2)perhronmachi
neI
u2=-
$(-
6/10)perhronmachi
neI
I
Max
Z=Mi
nZ*
=$96(
Tot
almi
nimum cost
)

64
 Not
e:Thedual
var
i esujar
abl eal edast
socal
l heshadowpr
ices
Z=300X1+250X2
g.Max.
Subj
ectt
o:
2X1+X2<40
X1+3X2<45
X1 <12
X1,X2>0
Thefinal
pri
mal
sol
uti
oni
s:
Cj 300 250 0 0 0

SV X1 X2 S1 S2 S3 Q

0 S1 0 0 1 -
1/3 -
5/3 5

250 X2 0 1 0 1/
3 -
1/3 11

300 X1 1 0 0 0 1 12

300 250 0 250/

3
Zj 6350
650/
3
Cj-Zj 0 0 0 -
250/
3 -
650/
3

Requi
red:
I
.Whatar
ethesol
uti
onoft
hedual
var
iabl
es(
i.
e.Shadowpr
ices)u1,u2andu3?
I
I.Whatist
heopt
imal
dual
cost
?
Solut
ion:
I
.Not
e: umns1cor
Thecol r ou1,
espondst s2t
ou2ands3t
ou3.
uj=-
(Cj-Zj)
Ther
efore,u1=0,u2=250/
3andu3=650/
3
I
I.MinZ* =$6350

Exer
cise
Wr
it
ethedual
oft
hef
oll
owi
ngpr
imal
LPpr
obl
ems
1.Mi
n.Z=3x1-
2x2+4x3
St
: 5.Mi
n.Z=8x1+3x2
3x1+5x2+4x3>7 St
:
6x1+x2+3x3>4 x1-6x2>2
7x1-
2x2-x3<10 5x1+7x2=- 4
4x1+7x2-2x3>2 x1,x2 >0
x1,x2,
x3>0 6.Mi
n.Z=3x1+x2+x3-
x4
St
:
Z=2x1+5x2+6x3
2.Max. 2x1-
x2+3x3+x4=1
x1+x2-x3+x4=3
x1,
x2, x4 >065
x3,
 St
:
5x1+6x2-
x3<3
2x1+x2+4x3<4
-
x1-5x2+3x3<1
3x1-
- 2x2+7x3<6
x1,x2,x3>0

3.Mi
n.Z=7x1+3x2+8x3
St
:
8x1+2x2+x3>3
3x1+6x2+4x3>4
4x1+x2+5x3>1
x1+5x2+2x3 >7
x1,x2,x3>0
Z=2x1+3x2+x3
4.Max.
St
:
4x1+3x2+x3=6
x1+2x2+5x3=4
x1,x2,x3>0

ASSI
GNMENTI
II
1. Amanuf
act
ureroft
oysmakest
wot
ypesoft
oys:
AandB.Pr
ocessi
ngoft
hese
t
wot
oysi
sdoneont nesX and Y.Thet
womachi oyA r
equi
rest
wohour
son

66
 neXandsi
machi xhour neY.ToyBr
sonmachi equi
resf
our neXandf
sonmachi ive
hour neY.her
sonmachi ear
esi
xteenhour
soft
imeperdayav
ail
abl neX
eonmachi
andt
hir
tyhour neY.Thepr
sonmachi ofi
tobt
ainedonbot
hthet
oysi
ssame,
i.
e.$5
pert
oy.
a.Whatshoul
dbet
hedai
l
ypr
oduct
ionofeachoft
het
wot
oys?
b.For
mul
atet
hepr
imal
andt LPPmodel
hedual
c. Fi
ndt
heopt
imal
sol
uti
onf
ort LPP
hedual
d.Rel
atet
hedual
opt
imal
sol
uti
ont
othepr
imal
opt
imal
sol
uti
on
e. Gi
vet
hepr
operi
nter
pret
ati
ont
othepr
imal
andt
hedual
sol
uti
ons.
2. oducesP,
Acompanypr QandRf
rom t
hreer
awmat
eri
alsA,
BandC.Oneuni
t
ofPr
equi
res2uni
tsofAand3ofB.
ProductQr
equi
res2uni
tsofBand5ofC.one
tofRr
uni equi
res3uni
tsofA,
2uni
tsofBand4uni
tsofC.Companyhas8uni
tsof
mat
eralA,
i 10uni
tsofmat
eralBand15uni
i tsofmat
eralCav
i ail
abl
etoi
t.Pr
ofi
tper
uni
tofpr sP,
oduct QandRar
e$3,
$5and$4r
espect
ivel
y.
a.For
mul
atet
hepr
obl
em mat
hemat
ical
l
y
b.Howmanyuni
tsofeachpr
oductshoul
dbepr
oducedt
omaxi
mizepr
ofi
t?
c. Wr
it
etheDUALpr
obl
em
3. The3rdandf
inal
simpl
ext
abl
eauf
ort
heLPPi
s:
Z=200x1+200x2
Max.
St
:
2x1+x2<8
x1+3x2<9
x1,x2>0
Cj $200 $200 $0 $0

SV X1 X2 S1 S2 Q
$200 X1 1 0 3/
5 -
1/5 3

250 X2 0 1 -
1/5 2/
5 2

Zj $200 $200 $80 $40 $1,

000
Cj-Zj 0 0 -
$80 -
$40
Whatar
ethesol
uti
onsoft
hedual
var
iabl u1,
es, u2andu3?Whati
stheopt
imal
dual
cost
?

CHAPTERV
DI
STRI
BUTI
ONMODELS
Onei
mpor
tantappl
i
cat
ionofl
i
nearpr
ogr
ammi
nghasbeeni
nthear
eaoft
hephy
sical

67
di
str
ibut
ion(
transpor
tat
ion)ofr
esour
ces,f
rom onepl
acet
o anot
her
,to meeta
speci
fi
csetofr
equi
rement
.

Thi
schapt
erdescr
ibest
wospeci
al–pur
poseal
gor
it het
hms:t ranspor
tat
ionmodel
heassi
andt gnmentmodel
.Modelf
ormul
ati
onandmanualsol
uti
onar
ecov
eredf
or
eachoft
hesecl
assesofpr
obl
ems.

ht
Bot r
anspor
tat
ionandassi
gnmentpr
obl
emsar
emember
sofacat
egor
yofl
i
near
pr
ogr
ammi
ngt
echni ednet
quescal
l wor
kfl
owpr
obl
ems.

5.
1.TRANSPORTATI
ONPROBLEMS
Tr
anspor
tat
ionpr
obl
em deal
swi
tht
hedi
str
ibut
ionofgoodsf
rom sev
eralpoi
ntsof
suppl
i
es(
sour
ces)t
oanumberofpoi
ntsofdemands(
dest
inat
ions)
.

Consi
deracor
por
ati
onengagedi
nthemanuf
act
ureofpr
oduct
s.Mostofsuchbi
g
cor
por
ati
ons ar
e of “
mul
ti
ple-
product
” and “
mul
ti
-uni
t” or
gani
zat
ions hav
ing
pr
oduct
ionuni
tssi
tuat
edatdi
ff
erentpl
aces.I
temsar
epr
oducedf
orsal
es.Sal
est
ake
pl
aceatdi
ff
erentmar
ket
swhi
char
e,agai
nlocat
edatdi
ff
erentpl
aces.I
tisnot
f
easi
blet
oco-
locat
epr
oduct
ionandmar
ket
.Mar
ket
sar
elocat
edawayf
rom t
he
manuf
act
uri
ngpl
aces.Hencepr
oduct
sar
esentt
ofact
orywar
ehousessetupnear
mar
ketout
l s.Costofpr
et oductconsi
stsofpr
oduct
ioncostanddi
str
ibut
ioncost
.

Costofpr
oduct=pr
oduct
ioncost+di
str
ibut
ioncost

Di
str
ibut
ioncostconsi
stsofmai
nlyt
het
ranspor
tat
ion costofi
temsf
rom i
ts
pr
oduct
ion(
manuf
act
uri
ng)cent
ert
othewar
ehouses.Tr
anspor
tat
iont
echni
quesar
e
desi
gned t
o mi
nimi
ze t
he di
str
ibut
ion cost
s.I
n or
dert
oident
if
ypr
oduct
s,i
tis
necessar
ytowor
koutperuni
tdi
str
ibut
ioncostofeachpr
oduct
.Weal
soknow t
he
pr
oduct
ioncapaci
tyofeachpr
oducti
neachf
act
oryi
snotf
ixed.Thehol
dingcapaci
ty
ofawar
ehouseorpot
ent
ialsal
esi
neachmar
ket
ingcent
eri
sagai
naf
ixedqual
i
ty
whi
chcannotbeexceeded.

Thechar
act
eri
sti
csoft
ranspor
tat
ionpr
obl
em ar
easf
oll
ows:
1.Al
i
mit
edsuppl
yofonecommodi
tyi
savai
l
abl
eatcer
tai
nsour
cesoror
igi
ns.
2.Ther
eisademandf
ort
hecommodi
tyatsever
aldest
inat
ions
3.Thequant
it
iesofsuppl
yateachsour
ceandt
hedemandateachdest
inat
ionar
e
const
ant.
4.Theshippi
ngort
ranspor
tat
ioncost
speruni
tfr
om eachsour
cet
oeachdest
inat
ion

68
 areassumedt obeconst ant.
5.Noshi pmentsareallowedbet weensourcesorbetweendest inati
ons.Allsupply
anddemandquant it
iesar egiveninwhol
enumberori ntegers.
6.Thepr obl
em istodet er
mi nehow manyuni tsshi
ppedf rom eachsour cetoeach
dest
inati
onsothatalldemandsar esati
sfiedatt
hemi ni
mum t otalshi
ppingcosts.

 Usesoft
ranspor
tat
iont
echni
ques:
1.Reducedist
ri
buti
onortransportat
ioncost
2.I
mpr ovecompeti
ti
venessofproduct
3.Assistpr
operl
ocati
onofwar ehouses
3.Assistpr
operl
ocati
onofnewf act
ori
esorplant
sbeingpl
anned.
4.Closedownwarehouseswhi charefoundcostl
yanduneconomi .
cal

 Theobj
ect
iveoft
ranspor
tat
ionpr
obl
em i
s:
1.Toi
denti
fytheopt
imalshi
ppingr
out es-
mini
mum costr out
e
2.Toi
denti
fythemaximum amountthatcanbeshippedovertheopti
mum rout
e
3.Todet
erminethet
otalt
ransf
ormationcostort
hepr of
itoftr
anspor
tat
ion

5.
1.1.The 
Transpor
tat
ion 
Met
hod
Thesol
uti
onal
gor
it
hm t
oat
ranspor
tat
ionpr
obl
em maybesummar
izedi
ntot
he
f
oll
owi
ngst
eps:

St
ep1:
For
mul
atet
hepr
obl
em andsetupi
nthemat
ri
xfor
m
Thef
ormul
ati
on oft
hepr
obl
em i
ssi
mil
art
othel
i
nearpr
ogr
ammi
ng.Her
ethe
obj
ect
ivef
unct
ioni
sthet
otalt
ranspor
tat
ioncostandt
heconst
rai
ntsar
ethesuppl
y
anddemandav
ail
abl
eateachsour
ceanddest
inat
ionr
espect
ivel
y.

St
ep2:
Obt
ainani
nit
ial
basi
cfeasi
blesol
uti
on

Ther e3met
ear hodst
ofi
ndt
hei
nit
ial
feasi
blesol
uti
on.

1.Nor
th-
WestCor
nerMet
hod(
NWCM)
2.LeastCostMet
hod(
LCM)
3.Vogel
’
sAppr
oxi
mat
ionMet
hod(
VAM)
Thei
nit
ialsol
uti
onobt
ainedbyanyoft
het
hreemet
hodsmustsat
isf
ythef
oll
owi
ng
condi
ti
on:

69
i
. Thesol
uti
onmustbef
easi
ble
i
.e.
:Itmustsat
isf
yal
lthesuppl
yanddemandconst
rai
nts

i
i
. Thenumberofposi
ti
veal
locat
ionsmustequalt
om+n-
1,wher
em=t
he
numberofr
ows(
oror
igi
nsorsuppl
ycent
ers)andn=t
henumberofcol
umns(
or
dest
inat
ioncent
ersordemandcent
ers)
Exampl
e:
m=3or
igi
nsandn=4dest
inat
ions==>m+n-
1=3+4-
1=6(
i.
e.t
het
ranspor
tat
ionmodel
shoul
dhav
e6occupi
edcel
l
s).

Not
e:
Thenumberofoccupi
edcel
ls<m+n-
1==>degener
atesol
uti
on

St
ep3:
Testt
hei
nit
ial
sol
uti
onf
oropt
imal
i
ty
I
fthe cur
rentsol
uti
on i
s opt
imal
,then st
op.Ot
her
wise,det
ermi
ne t
he new
i
mpr
oved 
sol
uti
on.
St
ep4:
Repeatst
ep3unt
ilanopt
imal
sol
uti
oni
sreached

5.
1.2.Li
nearpr
ogr
ammi
ngf
ormul
ati
onoft
het
ranspor
tat
ion
pr
oblem

Exampl
e
Supposet
hataf
ir
m hast
hreef
act
ori
es/
sour
cesofsuppl
y/&f
ourwar
ehouses
/
poi
ntofdemand/
.Thef
ir
m'spr
oduct
ioncapaci
tyatt
het
hreef
act
ori
es,t
hedemand
f
ort
hef
ourdi
str
ibut
ion cent
ersl
ocat
edatv
ari
ousr
egi
ons&t
hecostofshi
ppi
ng
eachuni
tfr
om t
hef
act
ori
est
othewar
ehousest
hrougheachr
out
eisgi
venasf
oll
ows:

Dest
inat
ions(
dd)=j

Or
igi
n Factory
(
Supply) W1 W2 W3 W4 Capacit
y=i
Br
.3 2 7 6
F1 5000
7 5 2 3 6000
F2

70
 2 5 4 5
F3 2500
Requi
rementsofthe
Warehouses 6000 4000 2000 1500 13500
(Uni
tsofdemand)

Usi
ngal
ltheabov
einf
ormat
ion
i
.Expr
esst
hef
oll
owi
ngt
ranspor
tat
ionpr
obl
em asanLPP
Sol
uti
on
Letxij=Theamountofcommodi
tyt
obet
ranspor
tedf
r cei(
om sour i=1,
2,
3)t
o
dest
inat
ionj
(j=1,
2,
3,
4).
Thent
heobj
ect
ivef
unct
ionoft
hepr
obl
em (
mini
mizat
ionoft
het
otalt
ranspor
tat
ion
cost
)canbef
ormul
atedas:

Mi
nZ=3x11+2x12+7x13+6x14+
7x21+5x22+2x23+3x24+
2x31+5x32+4x33+5x34
Subj
ectt
otheconst
rai
nts
a.Suppl
yconst
rai
nts:
x11+x12+x13+x14 =5000 F1suppl
yconst
rai
nt
x21+x22 +x23+x24=6000 F2suppl
yconst
rai
nt
x31+x32+x33+x34=2500 F3suppl
yconst
rai
nt
b.Demandconst
rai
nts:
x11+x21 +x31 =6000 W1 demandconst
rai
nt
x12+x22+x32 =4000 W2 demandconst
rai
nt
x13+x23 +x33=2000 W3 demandconst
rai
nt
x14+x24+x34 =1500 W4 demandconst
rai
nt
xij>0foral
li&j
I
ntheaboveLPP,t herearem xn=3x4=12deci si
onv ar
iabl
es&m +n=3+4=7
const
rai
nts.Thus,i
fthi
spr
obl
em i
ssol
vedbyt
hesi
mpl
exmet
hod,t
heni
tmayt
ake
consi
der
abl
ecomput
ati
onal
time.
i
i
.Thenet
wor
krepr
esent
ati
onoft
het
ranspor
tat
ionLPPi edNetwor
scal
l kfl
ow

Origi
n Desti
nati
oncenter
s
(SourcesofSuppl
y) (
Pointofdemandcent
ers)

F1 50000 3 W1 6000
2
6 7
7
5
71
F2 6000 W2 4000
2
3

5
W3 2000
2
4
F3 2500 5 W4 1500

Thi
sLPPhas12shi
ppi
ngr
out
es.Theobj
ect
ivei
stoi
dent
if
ythemi
nimum costr
out
e
(
Leastcostr
out
e).

5.
2.METHODSOFFI
NDI
NGI
NITI
ALFEASI
BLESOLUTI
ON
Ther
ear
esev
eralmet
hodsav
ail
abl
etoobt
ainani
nit
ialf
easi
blesol
uti
on.Her
ewe
shal
ldi
scussonl
ythr
eedi
ff
erentmet
hodst
oobt
aint
hei
nit
ial
feasi
blesol
uti
on:

A.NORTH-WESTCORNERMETHOD(
NWCM)
Thi
smet
hoddoesnott
akei
nt hecostoft
oaccountt ranspor
tat
iononanyr
out
eof
t
ranspor
tat
ion.

TheNWCM get
sit
snamebecauset
hest
art
ingpoi
ntf
ort
heal
l
ocat
ionpr
ocessi
sthe
UpperLeft
-hand(
Nor t
hwest
)cor
neroft
hetranspor
tat
iont
abl
e.Ther
efor
e,al
l
ocat
eto
heNor
t t
hwestcornerasmanyuni
tsaspossibl
e.

Nor
thwestcor
nerr
ule
Thef
oll
owi
ngsetofpr
inci
plesgui
dest
heal
l
ocat
ion:
1.Begi
nwi
t heupperl
ht efthandcel
l(Lef
t,uppermosti
nthet
abl
e),&al
l eas
ocat
manyuni
tsaspossi
blet
othatcel
l
.Thi
swi
l hesmal
lbet leramountofei
thert
he
r
ow suppl
yort
hecol
umndemand.Adj
ustt
her
ow &col
umnquant
it
iest
oref
lect
t
heal
l
ocat
ion.
2.Subt
ractf
rom t
her
owsuppl
y&f
rom t
hecol
umndemandt
heamountal
l
ocat
ed
3.I
fthecol
umndemandi
snow zer
o,mov
etot lnextt
hecel other
ight
,ift
her
ow
suppl
yiszer
o, edownt
mov othecel
li
nthenextr
ow.
I
fbot
har
ezer
o, ef
mov i
rstt
othenextcel
l her
ont ightt
hendownonecel
l
.
4.Onceacel
li
sident
if
iedasperst
ep(
3),
itbecomesanor
thwestcel
l.Al
l
ocat
etoi
tan
amountasperst 1)
ep(
5.Repeat
,theabov
est
eps(
1)-(
4)unt
ilal
lther
emai
ningsuppl
yanddemandi
sgone.
Exampl
e:
1)Consi
dert
hef
oll
owi
ngt
ranspor
tat
ionpr
obl
em:

To St
ore St
ore St
ore
St
ore2
1 3 4 Suppl
y
Fr
om

72
 Pl
ant1 19 30 50 10
7
70 30 40 60
Pl
ant2 9
40 8 70 20
Pl
ant3 18

5 8 7 14 34
Demand

a.Dev
elopani
nit
ial
feasi
blesol
uti
onusi
ngt
heNWCM
b.Comput
ethet
otal
costf
ort
hissol
uti
on.
Sol
uti
on
a. Tabl
e:I
nit
ialf
easi
blesol
uti
on
To
St
ore St
ore St
ore
St
ore2 Suppl
y
Fr
om 1 3 4
Pl
ant1 19 30 50 10
5 2 7
70 30 40 60
Pl
ant2 6 3 9
40 8 70 20
Pl
ant3 4 14 18

5 8 7 14 34
Demand

Checkt
hatt
hesol
uti
oni
sfeasi
bleornot
:
==>m +n-
1;m=3andn=4 3+4-
1=6cel
l
soccupi
ed(
Feasi
blesol
uti
on)

Thet
otal
transpor
tat
ioncostoft
hei
nit
ial
feasi
blesol
uti
onder
ivedbyt
heNWCM i
s:

Rout
e Uni
t Peruni
t Tot
al
Fr
om To Shi
pped X cost($) = Cost($)
Pl
ant1 St
ore1 5 19 95
pl
ant1 St
ore2 2 30 60
Pl
ant2 St
ore3 6 30 180
Pl
ant2 St
ore4 3 40 120
Pl
ant3 St
ore4 4 70 280
Pl
ant3 St
ore4 14 20 280
Tot
alCost
=$1015

Not
e:NWCM doesnotconsi
dert
hecostf
act
orf
oral
l
ocat
ion.

Exer
cise:

73
1.Det
ermi
neani
nit
ialbasi
cfeasi
blesol
uti
ont
othef
oll
owi
ngt
ranspor
tat
ionpr
obl
em
usi
ngNWCM.Comput
ethet
otal
costf
ort
hissol
uti
on

Dest
inat
ion
A B C Suppl
y
S1 2 7 14 5
S2
3 3 1 8
S3 5 4 7 7
S4 1 6 2 14
Demand 7 9 18

Answer
:X11=5,
X21=2,
X22=6,
X32=3,
X33=4,
X43=4,
andTot
alcost=$102

Not
e:
1.Tot
alSuppl
y=Tot
aldemand===>Bal
ancedTP
2.Tot
alSuppl
y≠t
ot demand===>Unbal
al ancedTP
3.
 Conv
ert
 t
he 
unbal
anced 
TP 
i
nto 
a bal
anced 
TP 
by 
usi
ng 
dummy
dest
inat
ion/
dummysour
ce.
*I
ftot
alSuppl
y>Tot
aldemand,
thencr
eat
eaf
ict
it
iousorar
ti
fi
cialdest
inat
ion
eddummydest
cal
l inat
ion
i
.e:
tot
alSuppl
y>Tot
aldemand===>Adddummycol
umn

*Excessdemand(
Suppl
y<demand)
-Addadummysour
ce
-Addadummyr
ow
Not
e:t shi
hecostof“ pment
s”t
othedummyi
susual
l
ysetatzer
o==>Nor
eal
cost

Exampl
e
Dev
elopani
nit
ial
feasi
blesol
uti
onusi
ngNWCM
Tabl
e:Unbal
ancedt
ranspor
tat
iont
abl
e
R S T Suppl
y
A 1 2 3 100
B
4 1 5 110

Demand 80 120 60 210

260

Sol
uti
on:

74
 R S T Suppl
y
1 2 3
A 100
80 20
4 1 5
B 110
100 10
0 0 0
Dummy 50
50
Demand 80 120 60 260

Answer
:X11=80,
X12=20,
X22=100,
X23=10,
X33=50Tot
alcost=$270

Assi
gnment
Consi hatHar
dert ley'
sSand&Gr
avelPi
thascont
ract
edt
opr
ovi
det
opsoi
lfort
hree
r
esi
dent
ialhousi
ng dev
elopment
s.Topsoi
lcan be suppl
i
ed f
orm t
hree di
ff
erent
“
far
ms"asf
oll
ows:
_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
__
Weekl
yCapaci
ty
Far
m (
CubicYar
ds)
A 100
B 200
C 200
_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
_

Demandf
ort
het
opsoi
lgener
atedbyt
heconst
ruct
ionpr
oject
sis:
_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
Weekl
yDemand

75
 Pr
oject (
Cubi
cYar
ds)
1 50
2 150
3 300
_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
__
Themanageroft
hesand&gr
avelpi
thasest
imat
edt
hecostpercubi
cyar
dtoshi
p
ov
ereachoft
hepossi
bler
out
es:
_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
__
Cost
spercubi
cyar
dto
Fr
om Pr
oject#1 Pr
oject#2 Pr
oject#3
Far
mA $4 $2 $8
Far
mB 5 1 9
Far
mC 7 6 3

_
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
__

Requi
red
Dev
elopt
hei
nit
ialf
easi
blesol
uti
onusi
ngNWCM &comput
ethet
otalcostf
ort
his
sol
uti
on.

B.THELEAST-COSTMETHOD(
LCM)or
(
LARGEST-PROFI
T)METHOD
LCM i
st hodusedami
hemet nimum costi
ntheal
l
ocat
ion.
I
tbegi
nsasol
uti
onbysequent
ial
l
yassi
gni
ngt
other
ati
osorcel
l
swi
t hemi
ht nimum
costasmanyunitsaspossibl
e.Thef
irstallocat
ionbemadet
othecel
lwi
tht
he
l
owestcost(
thehi
ghestpr
ofi
tinamaxi
mi zat
ioncase)
TheLeast
-CostMet
hodyi
eldsnotonl
yani
nit
ialf
easi
blesol
uti
onbutal
soonet
hat
i
scl
oset
oopt
imal
insmal
lpr
obl
ems.

Exampl
e
1.
Suppose t
hat a f
ir
m has t
hree f
act
ores / sour
i ces of suppl
y/& f
our
war
ehouses/
poi
ntofdemand/.
Thef
ir
m'spr
oduct
ioncapaci
tyatt
het
hreef
act
ori
es,
t
hedemandf
ort
hef
ourdest
inat
ioncent
ersl
ocat
edatv
ari
ousr
egi
ons&t
hecostof
shi
ppi
ngeachuni
tfr
om t
hef
act
ori
est
othewar
ehousest
hrougheachr
out
eisgi
ven
asf
oll
ows:
Dest
inat
ions
Factor
y
W1 W2 W3 W4 Capaci
ty

76
 F1 3 2 7 6 5000
F2 7 5 2 3 6000
F3 2 5 4 5 2500
Demand 6000 4000 2000 1500 13500

Requi
red:
a.Dev
elopani
nit
ial
feasi
blesol
uti
onusi
ngNWCM &Comput
ethet
otal
cost
b.Dev
elopani
nit
ialf
easi
blesol
uti
onusi
ngl
east
-costmet
hod&comput
ethet
otal
cost
.

Sol
uti
on:
I
nit
ial
feasi
blesol
uti
on
W1 W2 W3 W4 Fact
oryCapaci
ty
F1 3 2 7 6
5000
5000
Fact
ory 7 5 2 3
F2 6000
1000 4000 1000
2 5 4 5
F3 2500
1000 1500
Demand 6000 4000 2000 1500 13500

m=3,
n=4==>3+4-
1=6occupi
edcel
l
s(Feasi
ble)

Rout
es Uni
ts Uni
t Total
Fr
om To Shi
pped X Cost =Cost
F1 W1 5000 3 $15000
F2 W1 1000 7 7000
F2 W2 4000 5 20000
F2 W3 1000 2 2000
F3 W3 1000 4 4000
F3 W4 1500 5 7500
Tot
alt
ranspor
tat
ion =$55,
500
cost

b.
W1 W2 W3 W4 Fact
oryCapaci
ty

77
 3 2 7 6
F1 5000
1000 4000
7 5 2 3
F2 6000
2500 2000 1500
2 5 4 5
F3 2500
2500
Demand 6000 4000 2000 1500 13500

Rout
es Uni
ts Uni
t Total
Fr
om To Shi
pped X Cost =Cost
F1 W1 1000 3 $3000
F1 W2 4000 2 8000
F2 W1 2500 7 17500
F2 W3 2000 2 4000
F2 W4 1500 3 45000
F3 W1 2500 2 5000
Tot
alt
ranspor
tat
ion =$42,
000
cost

m=3,
n=4==>3+4-
1=6occupi
edcal
l
s(Feasi
ble)
Least
-Costmet
hodi
sbet
tert
hant
heNWCM becausei
tconsi
der
scostf
act
ori
es.

2.Devel
opt
hei
nit
ialf
easi
blesol
uti
onf
ort
hef
oll
owi
ngTP usi
ngt
hel
east
-cost
met
hod(
LCM)

Desti
nat
ion
D E F G Suppl
y
Source
A 1 5 3 4 100
B 4 2 2 5 60
C 3 1 2 4 120
demand 70 50 100 60 280

Sol
uti
on
The1stal
l
ocat
ionbemadet
othecel
lwi
tht
hel
east
-cost
.Cel
l
sAD&CEbot
hhav
ethe
l
owestcostf$1.Cel
lADi
ssel ed1stbecausemor
ect euni
tscanbeal
l
ocat
edt
oit(
70)
t
hant
ocel
lCE(
50)
.

Cel
lCFi
sfi
l
l n1stsi
edi nceal
argerquant
it
y(120-
50-
70)canbepl
acedt
her
e.Then,

78
t
her
emai
ningr
equi
rementof30f
orcol
umnFi
sal
l
ocat
edt
ocel
lBF&sour
ceB'
s
suppl
yisr
educedt
o30.

Thei
nit
ialsol
uti
onbyt
hel
east-
costmet
hod
D E F G Suppl
y
Sour
ce
A Dest
inat
ion 1 5 3 4 100
70
B 4 2 2 5 60
30 30
C 3 1 2 4 120
50 70
demand 70 50 100 60 280

m=3,
n=4==>3+4-
1=6occupi
edcel
l
s(f
easi
ble)

Rout
es Uni
ts Uni
t Total
Fr
om To Shi
pped X Cost =Cost
A D 70 1 $70
B F 30 2 60
B G 30 5 150
C E 50 1 50
C F 70 2 140
Tot
alt
ranspor
tat
ion =$470
cost

3.Devel
opani
nit
ial
feasi
blesol
uti
onusingLCM
R S T Suppl
y
A 1 2 3 100
B 4 1 5 110
Demand 80 120 60

Sol
uti
on

R S T Suppl
y
A 1 2 3 100
80 10 10
B 4 1 5 110
110
Dummy 0 0 0 50
50
Demand 80 120 60

79
Tot
al 
tr
anspor
tat
ion 
cost
=80x1+10x2+10x3+110x1+50x0=$240

Exer
cise
Thr
eegar
mentpl
ant
sar
eav
ail
abl
eformont
hlyeducat
ionoff
ourst
ylesofmen'
s
shi
rt
s.Thecapaci
ti
esoft
het
hreepl
ant
sar
e45,
000,93,
000and60,
000shi
rt
s.The
numberofshi
rt
srequi
redi
nst
yle"
a"t
hrough"
d"ar
e28,
000,
65,
000,
35,
000&70,
000,
r
espect
ivel
y.Thepr
ofi
ts,
in$pershi
rt
,ateachpl
antf
oreachst
ylear
eshownbel
ow.
Tabl
e:Thegar
mentpl
ant
s'pr
ofi
t.
STYLE a B C d
PLANT
1 8 12 -
2 6

2 13 4 3 10

3 0 7 11 8

Howmanyshi
rt
sofeacht
ypet
opr
oducei
neachpl
antsot
hatpr
ofi
tismaxi
mized?

Answer
s:Tot
alpr
ofi
t=$2,
119,
000
(
Hint
:Sel
ectt
hel
argestperuni
tpr
ofi
t,f
ir
st&usel
argest
-pr
ofi
tmet
hod)

2.Det
ermi
neani
nit
ial
feasi
blesol
uti
ont
othef
oll
owi
ngt
ranspor
tat
ionpr
obl
em
usi
ngLCM

A B C D Suppl
y
S1 1 5 3 3 34
Sour
ce
S2 3 3 1 2 15
S3 0 2 2 4 12
S4 2 7 2 4 19
demand 21 25 17 17

C.VOGEL'
SAPPROXIMATI
ONMETHOD(
VAM)
or
PENALTYMETHOD
VAM i
spr
efer
redt
otheot
hert
womet
hodsdescr
ibedabov
e.I
nthi
smet
hodeach
al
l
ocat
ioni
smadeont
hebasi
soft
heoppor
tuni
t orpenal
y( tyorext
ra)costt
hat
woul
dhav
eincur
redi
fal
l
ocat
ioni
ncer
tai
ncel
l
swi
thmi
nimum uni
ttr
anspor
tat
ion

80
costwer
emi
ssed.

I
nthi
smet
hod al
l
ocat
ionar
emadeso t
hatt
hepenal
tycosti
smi
nimi
zed.The
adv
ant
ageoft
hismet
hodi
sthati
tgi
vesani
nit
ialsol
uti
onwhi
chi
snear
ert
oan
opt
imal
sol
uti
onori
stheopt
imal
sol
uti
oni
tsel
f.

VAM det
ermi
nest
hepenal
t ornotusi
yf ngt
hemi
nimum costr
out
es,wher
ethe
obj
ect
ivei
stoav
oidl
argepenal
ti
essot
hatt
hepenal
tyf
rom notusi
ngt
her
out
esi
s
mi
nimi
zed.
Thest
epsi
nVAM ar
easf
oll
ows:
1.Cal
cul
atepenal
ti
esf
oreachr
ow (
col
umn)byt ng t
aki hesmal
lest& t
henext
smal
lestuni
ttr
anspor
tat
ioncosti
nthesamer
ow(
col
umn)
Thi
sdi
ff
erencei
ndi
cat
est
hepenal
tyorext
racostwhi
chhast
obepai
difone
f
ail
stoal
l
ocat
etot
hecel
lwi
tht
hemi
nimum uni
ttr
anspor
tat
ioncost

2.Sel
ectt
her
ow orcol
umnwi
tht
hel
argestpenal
ty&al
l
ocat
easmuchuni
tas
possi
blei
nthecel
lhav
ingt
hel
eastcosti
nthesel
ect
edr
oworcol
umnsat
isf
ying
t
hecondi
ti
ons.
I
fther
eisat
iei
nthev
aluesofpenal
ti
es,
thentcanbebr
okenbysel
ect
ingt
he
cel
lwher
emaxi
mum al
l
ocat
ioncanbemade.
3.Adj
ustt
hesuppl
y&demand&cr
ossoutt
hesat
isf
iedr
oworcol
umn
I
far
oworcol
umni
ssat
isf
iedsi
mul
taneousl
y,onl
yoneoft
hem i
scr
ossedout
&t
her
emai
ningr
ow (
col
umn)i
sassi
gnedazer
osuppl
y(demand).
Anyr
ow or
col
umnwi
thzer
o suppl
yordemandshoul
dnotbeusedi
ncomput
ingf
utur
e
penal
ti
es.
4.Repeatst
ep1t
o3unt
ilt
heent
ir
eav
ail
abl
esuppl
yatv
ari
oussour
ces&demandat
v
ari
ousdest
inat
ionsar
esat
isf
ied.

Exampl
e:
1. Det
ermi
ne an i
nit
ial basi
c f
easi
ble sol
uti
on t
o t
he f
oll
owi
ng

81
 t
ranspor
tat
ionpr
obl
em usingVAM.
Warehouse
A B C D Suppl
y Rowdiff
erenceorRowpenal
ty
oroppor
tunit
ycost
Fact
ory F1 2 2 0 4
25
5 20 2 0 - - -

F2 5 9 8 3
25 2 2 2 2 5
15 5 5
F3 6 4 3 2
10 1 2 2 - -
10
Demand 20 15 20 5 60

Columndiff
erence 3 2 3 1
orColumnpenalty
oropport
unit
ycost 3 2 - 1

1 5 - 1

5 9 - -

5 - - -
m=3,
n=4==>3+4-
1=6Occupi
edcel
l
s(f
easi
ble)
Thet
ranspor
tat
ioncostassoci
atedwi
tht
hissol
uti
oni
s:
Tot
alcost
=5x2+20x0+15x5x9=+95x3+10x4=$185
2.Adai
ryf
ir
m hast
hreepl
ant
slocat
edi
ndi
ff
erentr
egi
ons.Thedai
l
ymi
l
kpr
oduct
ion
ateachplantisasf oll
ows:
Pl
ant1:6mill
ionlit
er s.
Pl
ant2:1mi l
li
onl i
ters,&
Pl
ant3:10milli
onliters
Eachdayt hefir
m mustf ulf
il
ltheneedsofi
tsf
ourdi
str
ibut
ioncent
ers.Mi
nimum
r
equi
rementateachcenteri
sasfol
lows.
Dist
ri
buti
oncent
er1:7mil
li
onl
it
ers
" "2:5 " "
" "3:3 " "
" "4:2 " "

Costofshi
ppi
ngonemi
l
li
onl
i
ter
sfor
m eachpl
antt
oeachdi
str
ibut
ioncent
eri
s
gi
veni
nthef
oll
owi
ngt
abl
einhundr
edsofdol
l
ar.
Di
str
ibut
ionCent
er
D1 D2 D3 D4
P1 2 3 11 7
Pl
ant
P2 1 0 6 1
Fi
ndt
hei
nit
ialbasi
c
P3 5 8 15 9
f
easi
ble sol
uti
onby
:

82
 a.Nor
th-
westcor
ner
smet
hod
b.LCM
c.VAM i
ftheobj
ecti
stomi
nimi
zet
het
otal
transpor
tat
ioncost
Answer
:
a.Tot
alcost=$11,
600
b.Tot
alcost
=$11,
200
c.
D1 D2 D3 D4 Suppl
y RowPenal
ty
P1 2 3 11 7 6 1 1 5 -
P2 1 0 6 1 1 1 - - -
P3 5 8 15 9 10 3 3 4 4
Demand 7 5 3 2

Col
umn
Penal
ty
7 5 3 2
1 3 5 6
3 5 4 2
3 - 4 2
5 - 15 9

m+n-
1+3+4-
1=6==>t
hesol
uti
oni
snon-
degener
ate.
Thet
otal
transpor
tat
ioncost 1x2+5x3+1x1+6x5+3x15+1x9)x100=$10,
=$( 200

83
Assi
gnment
1.Det
ermi
neani
nit
ialbasi
cfeasi
blesol
uti
ont
othef
oll
owi
ngt
ranspor
tat
ionpr
obl
em
byVAM
Dest
inat
ion
D1 D2 D3 D4 Suppl
y
S1 21 16 15 3 11
Sour
ce S2 17 18 14 13 13
S3 32 27 18 41 19
Demand 6 10 12 15

2.Det
ermi
neani
nit
ial
feasi
blesol
uti
ont
othef
oll
owi
ngt
ranspor
tat
ionpr
obl
em usi
ng
a.NWCM
b.LCM,
&
c.VAM
Dest
inat
ion
D1 D2 D3 D4 Suppl
y
A 11 13 17 14 250
Sour
ce B 16 18 14 10 300
C 21 24 13 10 400
Deman 200 225 275 250
d

84
5.
3.OPTI
MALI
TYTESTS
Onceani
nit
ialsol
uti
oni
sav
ail
abl
e,t
henextst
epi
stochecki
tsopt
imal
i
ty.Anopt
imal
sol
uti
oni
sonei
nwhi
cht
her
eisnooppor
tuni
tycost
.Thati
s,t
her
eisnoot
hersetof
t
ranspor
tat
ionr
out
es(
all
ocat
ions)t
hatwi
l
lreducet
het
otal
oppor
tuni
tycost
.Thuswe
hav
etoev
aluat
eeachunoccupi
edcel
l(r
epr
esent
sunusedr
out
e)i
nthet
ranspor
tat
ion
t
abl
eint
ermsofoppor
tuni
tycost
.
Thepur
poseoft
heopt
imal
i
tyt
esti
stoseei
fthepr
oposedsol
uti
onj
ustgener
ated
canbei
mpr
ovedornot
.Thesol
uti
ont
obecheckedf
oropt
imal
i
tymustbenon-
degener
atei
. henoofoccupi
e.t edcel
lsmustbem+n-
1.
ThePr
ocedur
efort
est
ingopt
imal
i
tyi
sanal
ogoust
othatoft
hesi
mpl
exmet
hod.A
di
sti
nct
ioni weenbasi
smadebet cvar
iabl
es,t
hoseassoci
atedwi
thoccupi
edcel
ls&
non-
basi
cvar
iabl
es,
thoseassoci
atedwi
t heempt
ht ycel
ls.
Foreachempt
ycel
l
,theef
fectofchangi
ngi
ttoanoccupi
edcel
lisexami
ned.I
fany
oft
hesechangesar
efav
orabl
e,t
hesol
uti
oni
snotopt
imal&anewsol
uti
onmustbe
desi
gned.A f
avor
abl
e change meansan i
ncr
ease i
nthe v
alue oft
he obj
ect
ive
f
unct
ioni
nmaxi
mizat
ionpr
obl
emsoradecr
easei
nmi
nimi
zat
ionpr
obl
ems.

Opt
imum sol
uti
ont
oaTPcanbeobt
ainedbyf
oll
owi
ngt
womet
hods.Thesemet
hods
ar
emuchsi
mpl
ercompar
edt
osi
mpl
exmet
hodofanLPP.
A.St
eppi
ngst
oneMet
hod
B.Modi
fi
edDi
str
ibut
ionmet
hod(
MODIMet
hod)

A.St
eppi
ng-
stonemet
hod
TheSt
eppi
ng-
stonemet
hod i
sani
ter
ati
vet
echni
quef
ormov
ingf
rom ani
nit
ial
f
easi
blesol
uti
ont
oanopt
imal
sol
uti
oni
ntr
anspor
tat
ionpr
obl
ems.

hest
Fort eppi
ng-st
onemet
hodt
obeappl
i
edt
oat
ranspor
tat
ionpr
obl
em,oner
ule
henoofshi
aboutt ppi
ngr
out
esbei
ngusedmustbeobser
ved.Ther
ulei
s:

TheNo ofoccupi
“ edr
out
es(
orsquar
es)mustal
waysbeequalt
oonel
esst
hant
he
henoofr
sum oft owspl henoofcol
ust umns.
"
i
.e Occupi
ed shi
ppi
ng r
out es)=No ofr
es(squar ows+ No ofcol
umns-Non
degener
atesol
uti
on.

Rul
esf
ordr
awi
ngeachcl
osedl
oop:
1.Sel
ectanunusedsquar
e(cel
l
)tobeeval
uat
es.

85
 Begi
2. nni
ngatt
hiscel
l
,tr
aceacl
osedl
oopgoi
ngcl
ockwi
sedr
aw anar
row t
oan
occupi
edcel
li
nthesamer
ow(orcol
umn)
.

3.Mov
ever
ti
cal
lyorhor
izont
all
y(butneverdi
agonal
ly)t
oanot
heroccupi
edcel
l
“steppi
ng–over”unoccupi
edoroccupi
edcells(ifnecessar
y)withoutchangi
ng
them.Follow t
hesamepr ocedur
etootheroccupiedcell
sunti
lreturni
ngtothe
origi
nalemptycel
l
.
4.Begi
nwi
thapl
us(
+)si
gnatt
heunusedcel
l
,pl
aceal
ter
nat
ive(
-)si
gnsandpl
us
si
gnsoneachcor
nersquareofthecl
osedpat
hjustt
raced.
i
.eAteachtur
noft heloop(theloopmaycr ossoverit
sel
fatt
imes),pl
usand
minussi
gnsareal
ternat
elypl
acedinthecel
l
s,star
ti
ngwitha+signinanempty
cel
l.
5.Ther
emustbeexact
lyonecel
lwi
tha+si
gnandexact
lyonecel
lwi
tha-si
gni
n
anyr
oworcol
umni
nwhi
cht
hel
oopt
urns.
enno ofatl
6.Anev eastf
ourcel
lsmustpar
ti
cipat
einal
oopandt
heoccupi
edcel
ls
canbevi
sit
edonceandonl
yonce.
7.Repeatst
eps1t
o4unt
ilani
mpr
ovementi
ndexhasbeencal
cul
atedf
oral
lunused
squares(cel
l
s).I
falli
ndi
cescomputedar
egreat
ert
hanorequalt
ozer o,
anopti
mal
sol
utionhasbeenreached.I
fnot
,iti
spossi
blet
oimpr
ovethecurrentsol
uti
onand
decreasetot
alshi
ppingcost
s.

Not
e:
I
nanon-
degener
atepr
obl
em,t
her
eisonl
yonepossi
blewayofdr
awi
ngt
hel
oopf
or
eachempt
ycel
l
.

Howt
ofi
ndt
heval
ueofacel
leval
uat
or
Thev
alueofacel
lev
aluat
ori
sthesum oft
heperuni
tshi
ppi
ngcost
sint
hegai
ning
cel
l
slesst
hesum oft
heperuni
tshi
ppi
ngcost
sint
hel
osi
ngcel
l
soft
hecl
osedl
oop.
Thi
sev
aluat
ionpr
ocessmustbeext oal
endedt lunoccupi
edcel
l
s.

1.Thet
estofopt
imal
it
yforami
nimi
zat
ion(
cost
)pr
obl
em:
I
foneormor
eoft
hecel
lev
aluat
or snegat
si ive,
theexi
sti
ngsol
uti snotopt
oni imal
.
i
.e:
Formi
nimi
zat
ion(
cost
)pr
obl
ems,
all
thecel
lev
aluat
orsmustbeposi
ti
vef
or
opt
imal
i
ty.

Anal
ysi
softest
:
Checkal
ltheemptycel
l
sandsel
ectf
ori
mpr
ovementt
heonewi
tht
he
l
argesti
mprovementpot
enti
al.

86
 I
fthesol
uti snotopt
oni imal
,thenextst
epi
nthet
ranspor
tat
ionmet
hodi
sto
f
indabet
tersol
uti
on.Theoper
ati
onsi
nthi
sst
epar
e:
a.I
dent
if
yt i
he“ncomi
ng"cel
l(t
heempt
ycel
ltobeoccupi
ed)
I
n a mi
nimi
zat
ion case,t
he i
ncomi
ng cel
lisl
ocat
ed by
i
dent
if
yi hemostnegat
ngt ivecel
leval
uat
or.
b.Desi
gnani
mpr
ovedsol
uti
on
-
Byshi
ft
inguni
tsf
orm cel
ltocel
l

2.Thet
estofopt
imal
it
yforamaxi
mizat
ion(
prof
it
)case:
:
Ist
her
ever
sedofmi
nimi
zat
ioncase
-
Ifoneormor
eoft
hecel
lev
aluat
or sposi
si ti
ve,t
heexi
sti
ngsol
uti snot
oni
opt
imal
.
i
.e:f
oramaxi
mizat
ion(
prof
it
)case,al
lthecel
lev
aluat
orsmustbenegat
ive
f
oropt
imal
i
ty.I
fanycel
lev
aluat
i sposi
oni ti
ve,
thesol
uti
oni
snotopt
imal
.

Not
e:
 Acel
lev
al orof0i
uat ndi
cat
est
heexi
stenceofanot
hersol
uti
onj
ustasgood
ast
hecur
rentsol
uti
on.Thus,
int
hef
inalsol
uti
on,
ifcel
lev
aluat
orsof0exi
st,
t
hisi
ndi
cat
est
heex
i enceofmul
st ti
pleopt
imalsol
uti
ons.
 I
ftwoormor
ecel
l
shav
ethesamev
alue,
thenei
thermaybesel
ect
ed.
 I
ftwoormor
eoft
he"
losi
ng"cel
l
scont
ai hesameno ofuni
nt ts,bot
hwi
l
l
becomeempt
ysi
mul
taneousl degener
yanda“ ate"sol
uti
onwi
l
lresul
t.
 Fort
hemi
nimi
zat
ioncase;
whenoneormor
ecel
lev
aluat
orsar
enegat
ives,
the
cel
lwi
t hel
ht argestnegat
iveshoul
dbebr
oughti
ntosol
uti
onbecauset
hat
r
out hel
ehast argestpot
ent
ialf
ori
mpr
ovementperuni
t.
 The l
oop st
art
s and ends att
he sel
ect
ed unoccupi
ed cel
l
.Ev
ery cor
ner
el
ementoft oopmustbeanoccupi
hel edcel
l.

Exampl
e:
1.UseNWCM t
ofi
ndi
nit
ial
feasi
blesol
uti
onandt
estt
hesol
uti
onf
oropt
imal
i
ty.
Pr
ojec Pr
ojec Pr
ojec ss
t t t
A B C
F1 4 2 8 10
0
m
Far

F2 5 1 9 20
0
F3 7 6 3 20
0

87
 d 50 150 300 50
d
0

Sol
uti
on
I
nit
ial
feasi
blesol
uti
on

Pr
oject Pr
oject Pr
oject ss
A B C
F1 4 2 8 100
50 50
F2 5 1 9 200
100 100
F3 7 6 3 200
200
dd 50 150 300 500

m=3,
n=3==>3+3-
1=5(
Non-
degener
atesol
uti
on)
Tot
alt
ranspor
tat
ioncost=$[
50x4+50x2+100x1+100x9+200x3]=$1900

Pr
oject Pr
oject Pr
oject ss
A B C
F1 4 2 8 100 Ev
aluat
ion pat
h f
or t
he
50 50- + unoccupi
edcel
l
s
F2 5 1 9 200
100+ 100-
F3 7 6 3 200
200
Dd 50 150 300 500

Tabl
e:Testofopt
imal
i
ty
Unoccupi
edcel
ls Cel
leval
uat
ors
(
F2,
A) +5- 4+2-1=+2
(
F1,
C) +8-9+1-2=-2
(
F3,
A) +7-4+2-1+9-3=+10
(
F3,
B) +6-1+9-3=+11

88
Thenegat
ivev
aluef l(
orcel F1,C)i
ndi
cat
esani
mpr
ovedsol
uti
oni
spossi
ble.Foreach
uni
twecanshi
fti
ntot
hatcel
l
,thet
otalcostwi
l
ldecr
easeby$2.Thenextquest
ioni
s
howmanyuni
tscanber
eal
l
ocat
edi
ntot
hatcel
lwhi
l
eret
aini
ngt
hebal
anceofsuppl
y
anddemandf
ort
hatt
abl
e?

TheStepping-st
onepat
hf orcell(
F1,
C)i
s:
The+Si gnsinthepathindicat
eunitst he-si
obeadded,t gnsi
ndi
cateuni
tstobe
subt
ract
ed.Thel i
mitonsubt r
acti
oni hesmal
st lestquant
it
yinanegati
veposi
ti
on
al
ongt
hecel
lpat
h. Ther
ear
etwoquant
it
iesi
nnegat
iveposi
ti
ons,50and100.
Because50i
sthesmal
l
erquant
it
y,t
hatamountwi
l
lbeshi
ft
edi
nthef
oll
owi
ng
manner
:Subt
ract50uni
tsf
rom eachcel
lont
hepat
hwi
tha-si
gnandadd50uni
tst
o
t
hequant
it
yofeachcel
lwi
tha+si
gni
nit
.
Wi
theachi
ter
ati
on(
newsol
uti
on)
,iti
snecessar
ytoev
aluat
etheempt
ycel
l
stoseei
f
f
urt
heri
mpr
ovement
sispossi
ble.

Thedi
str
ibut
ionpl
anaf
terr
eal
l
ocat
ionof50uni
tsi
s:
A B C ss
F1 4 2 8 100
50 50
F2 5 1 9 200
150 50

F3 7 6 3 200
200
dd 50 150 300 500
Tabl
e:Testofopt
imal
i
ty
Unoccupi
edcel
ls Cel
leval
uat
ors
(
F1,
B) +2-
8+9-
1=+2

(
F2,
A) +5-
4+8-
9=0

(
F3,
A) +7-
4+8-
3 =+8

(
F3,
A) +6-
1+9-
3=+11

heseno i
Becausenoneoft snegat
ive,t
hisi
sanopt
imalsol
uti
on.Ther
efor
e,t
het
otal
costf
ort
hedi
str
ibut
ionpl
ani
s:
Thet
otal
transpor
tat
i 50x4+50x8150x1+50x9+200x3)=$1,
oncost=$( 800
2.Consi
dert
hef
oll
owi
ngTP
Dest
inat
ion
R S T ss
n
gi

A 1 2 3 100
i
Or

B 4 1 5 110

89
 210
dd 80 120 60
260
a.Obt
aint
hebasi
cfeasi
blesol
uti
onusi
ngVAM
b.Obt
aint
heopt
imal
sol
uti
on
c.Whati
stheopt
imal
shi
ppi
ngcost
?
Sol
uti
on:
a.I
nit
ial
feasi
blesol
uti
on
To
R S T ss
Oppor
tuni
tycost
1 2 3 100
A 1 1 1 1
80 10 10
om

4 1 5 110
Fr

B 3 3 - -
110
Dummy 0 0 0 50 0 - - -
dd 80 120 60 260
Oppor
tuni
ty 1 1 3
3 1 2
cost 1 2 3
1 2 -
Not
e:I
ncl
udet
hedummycel
l
stosel
ectt
heoppor
tuni
tycostunderVAM pr
obl
ems.

b.Testofopt
imal
it
y.
Tabl
e:Testofopt
imal
i
ty
Unoccupi
edcel
ls Cel
leval
uat
ors

+4-
1+2-
1=+4
(
B,R)
+5-
1+2-
3=+3
(
B,T)
+0-
1+3-
0=+2
(
D,R)
+0-
2+3-
0=+1
(
D,S)
Si
ncenoneoft
hecel
lev
aluat
orsi
snegat
ive,
theabov
efeasi
blesol
uti
oni
sopt
imal
.
Thus,
accor
dingl
ythedi
str
ibut
ioni
sasf
oll
ows
ASuppl
ies80uni
tstowarehouseR
BSuppl
ies10uni
tstowarehouseS
CSuppl
ies10uni
tstowarehouseT
BSuppl
ies110uni
tstowarehouseS
c.Thet
otal
opt
imal
shi
ppi s=$240
ngcosti

Exer
cise:

90
Consi
dert
hef
oll
owi
ngt
ranspor
tat
ionpr
obl
em

St
ore St
ore St
ore
Suppl
y
1 2 3

A 12 20 15 50
ehouses
B 9 11 4 15
war

C 20 14 8 55

Demand 25 50 45 120

a.Dev
elopani
nit
ialf
easi
blesol
uti
onusi
ngt
heNWCM.Andcomput
ethet
otal
costf
ort
hissol
uti
on.
b.Ev
aluat
ethesol
uti
onusi
ngt
hest
eppi
ng-
stonemet
hod.I
sthesol
uti
onopt
imal
?
Expl
ain.
c. Whati
sthet
otal
costf
ort
heopt
imal
sol
uti
on?

B.Modi
fi
edDi
str
ibut
ionMet
hod( hod)
MODIMet
I
tisanot
herl
ogar
it
hm f
orf
indi
ngt
heopt
imal
sol
uti
ont
oat
ranspor
tat
ionpr
obl
em.
TheMODImet
hodal
l
owsust
ocomput
eimpr
ovementi
ndi
cesqui
ckl
yfoeeach
unusedcel
lwi
thoutdr
awi
ngal
loft
hecl
osedpat
hs.Becauseoft
his,i
tcanof
ten
pr
ovi
de consi
der
abl
eti
me sav
ings ov
ert
he st
eppi
ng-
stone met
hod f
orsol
vi
ng
t
ranspor
tat
ionpr
obl
ems.
[

MODIpr
ovi
desanew meansoff
indi
ngt
heunusedr
out
ewi
tht
hel
argestnegat
ive
i
mpr
ovementi
ndex
.Oncet
hel
argesti
ndexi
sident
if
ied,
wear
erequi
redt
otr
aceonl
y
one cl
osed pat
h.Justaswi
tht
he st
eppi
ng-
stone appr
oach,t
hispat
h hel
pst
o
det
ermi
net mum Noofuni
hemaxi tst
hatcanbeshi
ppedv
iat
hebestunusedr
out
e.

St
epsi
ntheMODIMet
hod
Thest
epst
oev
aluat
eunoccupi
edcal
l
sar
easf
oll
ows:
1.Forani
nit
ialbasi
cfeasi
blesol
uti
on,cal
cul
ateUi andVj;
forr
owsand
col
umnsandset

Cij=Ui+Vjf
oral
loccupi
edcal
ls(
i,j
)
i
. Cel
e: lcost
=RawI
ndex+Col
umnI
ndex

2.Forunoccupi
edcel
ls,
cal
cul
ateoppor
tuni
tycostbyusi
ngt
her
elat
ion:

Kij=Cij–(
Ui+Vj)
;foral
ljandj.
WhereKijist
hecelleval
uatororoppor
tuni
tycost

3.Exami
net gnofeachKij
hesi

91
 Formi
nimi
zat
ioncase:
i
.IfKij>0,
thencur
rentbasi
cfeasi
blesol
uti
oni
sopt
imal
.
i
i
.IfKij =0,t
hent
hecur
rentbasi
cfeasi
blesol
uti
onwi
l
lremai
nbe
unaf
fect
edbutanal
ter
nat
ivesol
uti
onexi
sts.
i
i
i.I eKij<0,
foneormor thenani
mpr
ovedsol
uti
onbeobt
ainedent
eri
ng
unoccupi
edcell(i
,j)
,inthebasi
s.Anunoccupi
edcellhavi
ngthe
l
argestnegat
ivev
alueofKiji
schosenf
orent
eri
ngint
ot hesol
uti
on
mi
x(newt
ranspor
tat
ionschedul
e)
4.Sol
vet
hepr
obl
em asy
oudi
dusi
ngt
hest
eppi
ng-
stonemet
hod.
i
.e.const
ructacl
osedpat
h(orl
oop)f
ort
heunoccupi
edcel
lwi
thl
argest
negat
ive oppor
tuni
ty cost
.St
artt
he cl
ose pat
h wi
tht
he sel
ect
ed
unoccupi
edcel
landmar
kapl
ussi +)andi
gn( nthi
scel
l
,tr
aceapat
hal
ong
t
her
ows(
orcol
umns)t
oanoccupi
edcel
l
,mar
kthecor
nerwi
thmi
nussi
gn
-
()andcont
inuedownt
hecol
umn(
orr
ow)t
oanoccupi
edcel
landmar
k
t
hecor
nerwi
thpl
ussi +)andmi
gn( nussi -
gn()al
ter
nat
ivel
y.Cl
oset
he
pat
hbackt
othesel
ect
edunoccupi
edcal
l
.
Locat
ethesmal
l
estquant
it
yal
l
ocat
edt
oacel
lmar
kedwi
thami
nussi
gn.
Al
l
ocat
ethi
sval
uet
othesel
ect
edunoccupi
edcel
landaddi
ttoot
her
occupi
edcel
l
smar
kedwi
thpl
ussi
gnsandsubt
racti
tfr
om t
heoccupi
ed
cel
l
smar
kedwi
thmi
nussi
gns.

5.Obt
ainanewi
mpr
ovedsol
uti
onbyal
l
ocat
inguni
tst
otheunoccupi
edcal
l
andcal
cul
atet
henewt
ranspor
tat
ioncost
.

6.Testt
her
evi
sedsol
uti
onf
oropt
imal
i
ty.

Not
e:
Any i
nit
ialf
easi
ble sol
uti
on wi
l
ldo:NWCM,VAM Sol
uti
on,orany ar
bit
rar
y
assi
gnment
.
Thest
eppi
ng-st
onemet
hodi
sef
fi
cientf
orsmal
lsi
zedt
ranspor
tat
ionpr
obl
ems.
Forl
argerpr
obl
ems,
howev
er,
theMODImet
hodi
srecommended.

Exampl
e:
1.Obt
ainanopt
imalsol
uti
ont
othet
ranspor
tat
ionpr em byMODImet
obl hodgi
ven
bel
ow:

Pr
oject Pr
oject Pr
oject
o o Suppl
y
N1 N2 No3

Far
m1 4 2 8 100
Far
m2 5 1 9 200
Far
m3 7 6 3 200
50 150 300 500 92
Demand
Sol
uti
on
Not
e:
Bot
htheMODIandt
hest
eppi
ng-st
onemet
hodwi
l
lyi
eldst
hesamev
alues.
Remar
k:
Conv
ent
ional
l
y,webegi
nbyassi
gni
ngav
alueofzer
oast
hei
ndexf
orr U1=0)
ow1( .
Oncer
ow i
ndexhasbeenest
abl
i
shed,i
twi
l
lenabl
eust
ocomput
ecol
umni
ndex
number
sforal
loccupi
edcel
l
sint
hatr
ow.Si
mil
arl
y,onceacol
umni
ndexnumberhas
beendet
ermi
ned,i
ndexnumber
sforal
lrowscor
respondi
ngt
ooccupi
edcel
l
sint
hat
col
umncanbedet
ermi
ned.

Consi
dert
hei
nit
ial
feasi
blesol
uti
onoft
hegi
venexampl
ebyNWCM asshownbel
ow:

I
nit
ialsol
uti
on,
NWCM

Pr
oject Pr
oject Pr
oject
Suppl
y Ui
No1 No2 No3

Far
m1
4 2 8 100 U1=0
50

Far
m2
5 1 9 200 U2=1
100 100

Far
m3
7 6 3 200 U3=-
7
200

Demand 50 150 300 500

Vj V1=4 V2=2 V3=10

Todet
er neCij,
mi uset
heoccupi
edcel
l
s.

Fori
nst C11=4,
ance, C12=2,
C22=1,
C23=9,
andC33=3

Cij=Ui +Vj

93
 ==>C11=U1+V1==>4=0+V1==>V1=4 ,U1=0byconv
ent
ion
==>C12=U1+V2==>2=0+V2==>V1=2
==>C22=U2+V2==>1=U2+0==>U2=-1
==>C23=U2+V3==>9=-1+V3==>V3=10
==>C33=U3+V3==>3=U3+10==>U3=- 7

Not e:
Cij≠Ui +Vj(
Forunoccupi
edcel
ls)
Fori
nst
ance,
from t
heabov
einf
ormat
i C32≠U3+V2==>6≠-
on, 7+2

Tabl
e:Testofopt
imal
i
ty
Unoccupi
edcel
ls Cel leval
uat
ors
Kij=Cij–(
Ui+Vj)
(
1,3) C13–(
U1+V3)
=8-
(0+10)
=-2
(
2,1) C21–(
U2+V1)
=5-
(-
1+4)
=+2
(
3,1) C31–(U3+V1)=7-
(-7+4)=10
(
3,2) C32–(
U3+V2)=6-(
-7+2)=+11

I
nthi
scase,wef
oundhatcel
l(1,3)hadanev
aluat
ionof-
2,whi
chr
epr
esent
edan
i
mpr
ovementpot
ent
i ofand$2peruni
al t.Hence,
ani
mpr
ovedsol
uti
oni
spossi
ble.

Thest
eppi
ng-
stonepat
hforcal
l(1,
3)i
s:

Pr
oject Pr
oject Pr
oject
Suppl
y
No1 No2 No3

Far
m1
4 2 8 100
50 50

Far
m2
5 1 9 200
100 100

Far
m3
7 6 3 200
200

Demand 50 150 300 500

Thedi
str
ibut
ionpl
anaf
terr
eal
l
ocat
ionof50uni
tsi
s:

Pr
oject Pr
oject Pr
oject
Suppl
y Ui
No1 No2 No3

Far
m1
4 2 8 100 U1=0
50 50

Far
m2
5 1 9 200 U2=1
+100 100-
- +
Far
m3
7 6 3 200 U3=-
5
200

Demand 50 150 300 500

94

Vj V1=4 V2=0 V3=8

 Cij=Ui +Vj
==>C11=U1+V1==>4=0+V1==>V1=4 ,U1=0byconv
ent
ion
==>C13=U1+V3==>8=0+V3==>V3=8
==>C23=U2+V3==>1=U2+0==>U2=1
==>C22=U2+V2==>1=1+V2==>V2=0
==>C33=U3+V3==>3=U3+8==>U3=- 5
Table:Testofopti
mali
ty
Unoccupi
edcel
ls Cel leval
uat
ors
Kij=Cij–(
Ui+Vj)
(
1,2) C12–(
U1+V2)
=2-
(0+0)
=2
(
2,1) C21–(
U2+V1)
=5-
(1+4)
=0
(
3,1) C31–(U3+V1)
=7-(
-5+4)=8
(
3,2) C32–(U3+V2)
=6-(
-5+0)=+11

Becausenoneoft
hecel
lev
aluat
orsi
snegat
ive,
thi
sisanopt
imal
sol
uti
on.
Thus,
thet
otal
costf
ort
hedi
str
ibut
ionpl
an=$1800

95
Assi
gnment
1.Obt
ainanopt
imalsol
uti
ont
othet
ranspor
tat
ionpr em byMODImet
obl hodgi
ven
bel
ow.

D1 D2 D3 D3 Suppl
y

Far
m1
19 30 50 10 7

Far
m2
70 30 40 60 9

Far
m3
40 8 70 20 18

5 8 7 34
Demand 14

2.Acompanyhasfourwarehousesa,b,candd.I
tisr
equi
redt
odeli
veraproduct
f
rom t
hese war
ehousestothr
eecust sA,
omer BandC.Thewar
ehouseshavethe
f
oll
owi
ngamount
sinst
ock:

War
ehouse: a b c d
Noofuni
ts: 15 16 12 13

Andt
hecust
omer
’sr
equi
rement
sar
e:

96
 Cust
omer
: A B C
Noofuni
ts: 18 20 18

Thet
abl
ebel
ow showst
hecost
soft
ranspor
ti
ngoneuni
tfr
om war
ehouset
o
cust
omer
.

A b c d

8 9 6 3
A

6 11 5 10
B

3 8 7 9
C

Fi
ndt
heopt
imal
transpor
tat
ionr
out
es.

3.A manuf
actur
erhas dist
ri
buti
on cent
ers atX,Y and Z.These cent
ers hav
e
av
ail
abi
li
ty40,20,and40uni
tsofhisproduct
.Hi
sret
ailout
l satA,B,C,DandE
et
r
equi
re25,
10,
20,
30and15uni
tsr
espect
ivel
y.Thet
ranspor
tat
ioncost(
in$)peruni
t
bet
weeneachcent
reout
leti
sgi
venbel
ow:

Dist
ribut
ion
Ret
ailout
let
s
center

A B C D
E
X
55 30 40 5
500 40
55 30 40

35 30 100 45 60
Y

40 60 95 35 30
Z

Det
ermi
net
heopt
imal
dist
ri
but
iont
omi
nimi
zet
hecostoft
ranspor
tat
ion.

4.Acompanyhasf
act
oresA,BandC whi
i ehousestD,E,FandG.Mont
chwar hly
f
act
orycapaci
ti
esar
e160,
150and190uni
tsr
espect
ivel
y.Mont
hlywar
ehouses
r
equi
rement
sar
e80,
90,
110&160uni
tsr
espect
ivel
y.Uni
tshi
ppi
ngcost
s(i
n$)ar
e
asf
oll
ows:
To
D E F G

A 42 48 38 3
507
55 30 40
om

40 49 52 51
Fr

39 38 40 43
C 97
 Det
ermi
net
heopt
imadi
str
ibut
ionf
ort
hiscompanyt
omi
nimi
zeshi
ppi
ngcost
s.

5.
4.SPECI
ALCASES
5.
4.1.Degener
acy
Acondi
ti
ont
hatoccur heNo ofoccupi
swhent edcel
l
sinanysol
uti
onsl
esst
hant
he
Noofr
owspl heNoofcol
ust umnsmi
nus1i
nat
ranspor
tat
iont
abl
e.

i
.e.Noofoccupi
edcal
ls<m+n-
1..
..
..
..
..
..
..
..
..
..
.Degener
acy

I
ft he No ofoccupi
ed cel
ls = m+n-
1,t
hen t
he sol
uti
on i
s non-
degener
ate.
Thedegener
acyi
nthet
ranspor
tat
ionpr
obl
emsmayoccuratt
wost
ages:
1.whenobt ai
ninganinit
ialsol
uti
on
2.Duringimprovement( oratanystagewhi
lemovi
ngtowardsopt i
mal
soluti
on.
Toresol
vedegeneracy
,wepr ocessedbyal l
ocatngav
i erysmal
lquanti
tycloseto
zer
otooneormor eunoccupiedcellsoast ogetm+n-1numberofoccupi
edcells.
Thi
samounti
sdenot
edbyaGr
eekl
eter(
t epsi
l
on)or (
del
ta)
.Thi
squant
iywoul
t d
notaf
fectt
het
otalcostaswel
lassuppl
yanddemandval
ues.
=Al
mostzer
o
I
nami
nimi
zat
iont
ranspor
tat
ionpr
obl
em,al
l
ocat
ethesmal
l
estt
ranspor
tat
ioncost
o
t .
I
namaxi
miz
ati
ont
ranspor
tat
ionpr
obl
em,i
tshoul
dbeal
l
ocat
edt
oacel
lthathasa
hi
ghpayof
fval
ue.

I twheni
nser tisabl
etocr eacl
eat osedl
oopf
oreachoccupi
edcel
l
.

Thepur
poseofepsi
l
on/
del
tai
stoenabl
eev
aluat
ionoft
her
emai
ningempt
ycel
l
s.
Thechoi
ceofl
ocat
ionf
ort
heepsi
l
on/
del
tacanbesomewhatt
ri
cky
:someempt
y
cel
l
smaybeunsui
tabl
eift
heydonotenabl
eev
aluat
ionsofr
emai
ningempt
ycel
l
s.
Notal
lchoi
ceswoul
dbeaccept
abl
e.
Act
ual
l
y heNo ofepsi
,t l
on/
del
tasneededwi
l
lequalt
hedi
ff
erencebet heNo of
weent
compl
et sandm+n-
edcel
l 1.Hower
ver
;youwi
l
lonl
ybeexposedt
othemostcommon
casei
nwhi
chonemor
ecompl
etedcel
li
sneeded.
Theepsi
l
on/
del
tacannotbepl
acedi
nacel
lwhi
chl
atert
urnsoutt nanegat
obei ive

98
posi
ti
on ofacel lpat
hinvol
vedi
nr eal
l
ocat
ionbecauseepsi
lon/
del
tawillbet
he
“smal
lestquant
it
yanegati
veposi
ti
on“andshif
ti
ngthatminut
equant
ityar
oundt
he
cel
lpat
hwi
l
lleav
ethesol
uti
onv
irt
ual
l
yunchanged.Consequent
ly,acer
tai
namount
oft
ri
alander
rormaybenecessar
ybef
oreasat
isf
act
oryl
ocat
ioncanbei
dent
if
iedf
or
epsi
l
on/
del
ta.

Example
1.Sol
vet
hef
oll
owi
ngt
ranspor
tat
ionpr
obl
em.

1 2 Suppl
y
1 3 3
50
2 4 6 30
Demand 50 30 80
Sol
uti
on:
Usi
ngNWCM andMODI
,thei
nit
ial
sol
uti
oni
s:

Ui
1 2 Suppl
y

3 3
1
50
50 U1=0

4 6
2 30 U2=3
30
Demand 50 30 80
Vj V1=3 V2=3

Cij=Ui +Vj
==>C11=U1+V1==>3=0+V1==>V1=3,U1=0byconv
ent
ion
==>C12=U1+V2==>3=0+V2==>V2=3
==>C22=U2+V2==>6=U2+3==>U2=3
==>C33=U3+V3==>3=U3+8==>U3=-5
Not
e:m=2andn=2==>2+2-
1=3==>Occupi
edcel
l
s=2<3(
Degener
acy
)
Tabl
e:Testofopt
imal
i
ty
Unoccupi
edcel
l Celleval
uator
Kij=Cij–(
Ui+Vj)
(
2,1) C21–(
U2+V1)
=4-
(3+3)=-
2

Theopt
imal
sol
uti
oni
s:
Ui
1 2 Suppl
y

3 3 Cij=Ui +Vj
1 50 30 50 U1=0 ==>C11=U1+V1==>3=0+V1==>V1=3, U1=0 by
4 6
2 30 U2=1
30
99
Demand 50 30 80

Vj V1=3 V2=3
conv
ent
ion
==>C21=U2+V1==>4=U2+3==>U2=3
==>C12=U1+V2==>3=0+V2==>V2=3

Tabl
e:Testofopt
imal
i
ty
Unoccupi
edcel
ls Cel leval
uat
ors
Kij=Cij–(
Ui+Vj)
(
2,2) C22–(
U2+V2)
=6-
(1+3)
=+2

Thet
otal
cost 20x3+30x3+30x40=$270
=$(
5.
4.1.Al
ter
nat
iveOpt
imalsol
uti
ons
Theesi
st6aneofal
ter
nat
iveopt
imalsol
uti
oncanbedet
ermi
nedbyani
nspect
ionof
t
heoppor
tuni
tycost
s,Kijf
ort
heunoccupi
edcel
l
s.I
fanunoccupi
edcel
linanopt
imal
sol
uti
onhasoppor
tuni
tycostofzer
o,t
henanal
ter
nat
iveopt
imalsol
uti
oncanbe
f
ormedwi
thanot
hersetofal
l
ocat
ionswi
thouti
ncr
easi
ngt
het
otalt
ranspor
tat
ion
cost
.

Exer
cise
XYZTobaccoCompanypur
chasest
obaccoandst
oresi
nwar
ehousesl
ocat
edi
nthe
f
oll
owi
ngf
ourci
ti
es.

war
ehousel
ocat
ion Capaci
ty(
Tones)
Ci
tyA 90
Ci
tyB 50
Ci
tyC 80
ci
tyD 60

Thewar
ehousessuppl
ytobaccot
oci
gar
ett
ecompani
esi
nthr
eeci
ti
est
hathav
ethe
f
oll
owi
ngdemand:

Ci
gar
ett
eCompany Demand(
Tones)
L 120

100
 P 100
Q 110

The f
oll
owi
ng r
ail
road shi
ppi
ng cost
s pert
one (
in hundr
ed dol
l
ars)hav
e been
det
ermi
ned:

War
ehouse L p Q
l
ocat
ion
A 7 10 5

B 12 9 4

C 7 3 11

D 9 5 7

Becauseofr
ailr
oadconst
ruct
ion,shi
pment
sar
etempor
ari
l
yfr ehouseatci
om war tyA
oLCi
t gar
ett
eCompany.
A.Fi
ndt
heopt
imum di
str
ibut
i orXYZ TobaccoCompany
onf .
B.Ar
ether
emul
ti
pleopt
imum sol
uti
ons?I
fther
ear
eal
ter
nat
iveopt
imum
sol
uti
ons,
ident
if
ythem.

5.
4.3.Pr
ohi
bit
edTr
anspor
tat
ionRout
es.
Thesi
tuat
ionmayar
isesuchasr
oadhazar
ds(
snow,f
oods,et
c),t
raf
fi
cregul
ati
onet
c,
wheni
tisnotpossi
blet
otr
anspor
tgoodsf
rom cer
tai
nsour
cest
ocer
tai
ndest
inat
ions.
I
nthi
scase,t
heappr
opr
iat
ecel
lmayei
therbecompl
etel
ycr
ossedoutorav
eryl
arge
peruni
ttr
anspor
tat
ioncostassi
gnt
oi M)
t(

5.
5.ASSI
GNMENTPROBLEMS
TheAssi
gnmentPr
obl
em(
AP)r
efer
stot assofLPPst
hecl hati
nvol
vesdet
ermi
ning
t
hemostef
fi
cientassi
gnmentofpeopl
eto pr
oject
s,sal
espeopl
etot
err
it
ori
es,
cont
ract
stobi
dder
s,j
obst
omachi
nes,andsoon.Theobj
ect
ivei
stoassi
gna
numberofr
esour
cest
oanequalnumberofact
ivi
ti
essoast
omi
nimi
zet
otalcost
sor
t
otal
timeormaxi
mizet
otal
prof
itofal
l
ocat
ion.

The pr
obl
em ofassi
gnmentar
ises because av
ail
abl
eresour
ces such as men,
machi
nes,et
chav
evar
yingdegr
eesofef
fi
ciencyf
orper
for
mingdi
ff
erentact
ivi
ti
es
suchasj
ob.Ther
efor
e,cost
,pr
ofi
tort
imeofper
for
mingt
hedi
ff
erentact
ivi
ti
esi
s
di
ff
erent
.
Assumpt
ions:

101
TheAPi
saspeci
alcaseofTPundert
hecondi
ti
ont
hatt
henumberofor
igi
nsi
sequal
t
othenumberofdest
inat
i z.m=n.
ons.Vi Henceassi
gnmenti
smadeont
hebasi
sof
1:
1.

Fol
lowi
ngar
etheassumpt
ions:
 Numberofj
obsi
sequalt
othenumberofmachi
nesorper
sons
 Eachmanormachi
nei
sloadedwi
thoneandonl
yonej
ob.
 Eachmanormachi neisindependent
lycapableofhandli
nganyofthejobsbeing
presented.
 Loadingcr it
eri
amustbecl earl
yspeci
fiedsuchas“mi ni
mizi
ngoperat
ingtime”or
“maximi zingprofi
t”,
or“minimizi
ngproductioncost”or“mini
mizi
ngt hr
oughout
(productioncycle)t
ime”etc.

Remar k:
TheAPi sconsider
edasaspeci alTPinwhichthesuppl
yateachsour
ceandthe
demandateachdesti
nat
ionarealwaysoneunit
.
Sincethesuppl
yanddemandar ealwaysequalt
ooneuniti
neachrowandcol
umn,
ther
eisnoneedtowri
tethem i
nt heassi
gnmenttabl
e.
Exampl
e:
Ser
vicecost
sofdi
ff
erentt
eam assi
gnment(
$int
housands)
Tabl
e:Theassi
gnmentt
abl
e
Zone
Z1 Z2 Z3 Zone
Servi
ce Z1 Z2 Z3 Suppl
y
Team Servi
ce
S1 20 15 31 Team
====> S1 20 15 31 1
S2 17 16 33
S2 17 16 33 1
S3 18 19 27
S3 18 19 27 1
Deman
1 1 1
d

Theabov
epr
obl
em canbepr
esent
edasaLPPasf
oll
ows:
Mi
nZ=20x11+15x12+31x13+17x21+16x22+33x23+18x31+19x32+27x33
Subj
ectt
otheconst
rai
nts
a.Suppl
yconst
rai
nts:
x11+x12+x13=1 S1const
rai
nt
x21+x22 +x23=1 S2const
rai
nt
x31+x32+x33=1 S3suppl
yconst
rai
nt
b.Demandconst
rai
nts

102
 x11+x21 +x31 =1 Z1 const
rai
nt
x12+x22+x32 =1 Z2 const
rai
nt
x13+x23 +x33=1 Z3 constr
aint
xijei
ther0or1foral
li,j
Si
nceal
lxijcanbeei ther0or1,t
herewi ll
beoneassi
gnmenti
neachsuppl
yconst
rai
nt
andoneassi
gnmenti
neachdemandconst
rai
nt.

Asi
nthet
ranspor
tat
ionpr
obl
em,assi
gnmentpr
obl
emscanbebal
ancedornot
.Ina
bal
ancedcase,
thenumberofobj
ect
stobeassi
gnedequal
sthenumberofobj
ect
sto
whi
cht
heyar
eassi
gned.Unbal
ancedpr
obl
em canbebal
ancedbyaddi
ngadummy
(
dummi
es)wi
thzer
ocostcoef
fi
cient
s.

5.
5.1.Met
hodsofsol
vi
ngassi
gnmentpr
obl
ems
Anassi
gnmentpr
obl
em canbesol
vedbyt
hef
oll
owi
ngmet
hods:
1.Enumer
ati
onmet
hod
2.Si
mpl
exmet
hod
3.Tr
anspor
tat
ionmet
hod
4.Hungar
ianmet
hod

5.
5.1.
1.Hungar
ianMet
hod/
Flood’
sTechni
que/
The Hungar
ian Met
hod (
dev
eloped by Hungar
ian mat
hemat
i an D.
ci Koni
g) of
assi
gnmentpr
ovi
desuswi
thef
fi
cientmet
hodoff
indi
ngt
heopt
imalsol
uti
onwi
thout
hav
ingt
omakeadi
rectcompar
isonofev
erysol
uti
on.I ksont
twor hepr
inci
pleof
reduci
ng thegivencostmatr
ix(thepri
nci
pleofmat r
ixreduct
ion)toamatrixof
opport
uni
tycosts,whi
chmeansthatbysubt
ract
ingandaddingappropr
iat
enumbers
i
nt hecosttabl
eormat r
ix,
wecanr educetheprobl
em toamat r
ixofoppor
tuni
ty
costs.

Oppor
tuni
tycost
sshowt
her
elat
ivepenal
ti
esassoci
atedwi
thassi
gni
ngr
esour
cet
o
anact
ivi
tyasopposedt
omaki
ngt
hebestorl
east
-costassi
gnment
.Ifwecanr
educe
t
hecostmat
ri
xtot
heext
entofhav
ingatl
eastonezer
oineachr
owandcol
umn,
then
i
twi
l
lbepossi
blet
omakeopt
imal
assi
gnment
s.

St
epsi
nsol
vi
ngassi
gnmentpr
obl
ems:
St
ep.
1.Devel
opt
hecostt
abl
efr
om t
hegi
venpr
obl
em
I
fthenumberofr
owsdoesnotequalt
henumberofcol
umnsandv
icev
ersa,t
hena
dummyr
ow ordummycol
umnmustbeadded.Theassi
gnmentcost
sfordummy

103
cel
l
sar
eal
way
szer
o.

St
ep2.
Findt
heoppor
tuni
tycostt
abl
e
i
.e.Thet
ransf
ormat
ionoft
hecostmat
ri
xto whati
ster
med asat
otal
-
oppor
tuni
tycostmat
ri
x.
I
tinv
olv
est
wooper
ati
ons:
a.Per
for
mrowoper
ati
on
i
.e.Locat
ethesmal
l
estel
ementi
neachr
ow oft
hegi
vencostt
abl
e
andt
hensubt
ractt
hatt
hegi
vencostt
abl
eandt
hensubt
ractt
hatf
rom
eachel
ementoft
hatr
ow
b.Per
for
m col
umnoper
ati
on
i
.e.I
nther
educed mat
ri
xobt
ained f
rom 2(
a),
l
ocat
ethesmal
l
est
el
ementi
neachcol
umnandt
hensubt
ractt
hatf
rom eachel
ementof
t
hatcol
umn.Not
icet
hateachr
owandcol
umnnowhav
eatl
eastone
zer
oval
ue.
St
ep3.
Testf
oranopt
imalassi
gnment
i
.e.Testt
het
abl
eresul
ti
ngf
rom st
ep2t
oseewhet
heranopt
imalassi
gnmentcanbe
made.Thepr
ocedur
eis:
a.Drawt hemi nimum numberofHor
izont
aland/
orVer
ti
call
i
nesnecessar
yto
coverallzeroscosts.
Dr
aw t
hel
i
nesbyt
ri
alander
rorbutal
way
str
ytocov
ert
woormor
e
zer
oswi
thonel
i
ne.
b.Countt
henumberoft
hel
i
nes
I
fthenumberofl
i
nesequal
sei
thert
henumberofr
owsorcol
umnsi
n
t
het
abl
e,anopt
imal
assi
gnmentcanbemade.I
fthenumberofl
i
nesi
s
l
esst
hant
henumberofr
owsorcol
umns,
ani
mpr
ovementi
spossi
ble
(
wepr
oceedt
ost
ep4)
.

St
ep4.
Impr
ovet
hepr
esentoppor
tuni
tycostt
abl
e(mat
ri
x)
Thi
sisdonebyt
hef
oll
owi
ngoper
ati
ons:
a.Fi
ndt
hesmal
l
estent
ryi
ntheuncover
edcel
l
s(cel
l
swi
thnol
i
nest
hrought
hem)and
subtr
actitfrom al
lentri
esintheuncoveredcel
ls.
b.Addt hesamesmal lestentr
yt othosecell
sinwhicht heli
nesint
ersect(
cel
lswith
twoli
nest hem)
c.Cellswithonel inethroughthem aretransf
err
ed( i
.e.unchangedtotheimproved
table)
.

I
nthosepr
obl
emswher
ethef
isti
r mpr
ovementdoesnotyi
eldanopt
imal
sol
uti
on,
we
keepon i
mpr
ovi
ngt
hesol
uti
onbyr
epeat
ingst
ep4unt
ilanopt
imalsol
uti
oni
s

104
achi
eved.

St
ep5:Makeanopt
imalassi
gnment
Anopt
imalassi
gnmentshoul
dbemadet
ocel
l
swi
thazer
oent
ery
,mai
ntai
ningt
he
one-
to-
oner
equi
rement

I
fmor
ethanoneopt
imalsol
uti
onexi
sts,at
ri
al-
and–er
rorappr
oachcanbeusedt
o
f
indal
lpossi
blecombi
nat
ionassi
gnment
sint
hezer
ocel
l
s.

Not
ethatmul
ti
pleopt
imalsol
uti
onsar
epossi
ble.

Exampl
e:
1.Acomput
ercent
erhast
hreepr
ogr
ammer
s.Thecent
erwant
sthr
eeappl
i
cat
ion
pr
ogr
amst
obedev
eloped.Theheadoft
hecomput
ercent
er,af
terst
udy
ingcar
eful
l
y
t
hepr
ogr
amst
obedev
eloped,
est
imat
ethecomput
ert
imei
nmi
nut
esr
equi
redbyt
he
exper
tsf
ort
heappl
i
cat
ionpr
ogr
amsasf
oll
ows:

Programs
(
Est
imat
edtimeinmi
nut
e)
s
ammer

A B C
1 120 100 80
ogr

2 80 90 110
Pr

3 110 140 120

Assi
gnt
hepr
ogr
ammer
stot
hepr
ogr
amsi
nsuchawayt
hatt
het
otalcomput
ert
ime
i
smi
nimum.

Sol
uti
on:
St
eps1and2:
a.Per
for
mrowr
educt
ion
Themi
nimum t
imeel
ementi
nrow1,
2,and3i
s80,
80and110r
espect
ivel
y.Subt
ract
t
hoseel
ement
sfr
om al
lel
ement
sint
hererespecti
verow.Ther educedt
imemat
ri
xis:
Tabl
e:Aft
errowreduct
ion

A B C
-
80 1 40 20 0
-
80 2 0 10 30
-
110 3 0 30 10

b.Col
umnr
educt
ion

105
Si
ncecol
umnBhasnoone‘
0’
,per
for
m al
socol
umnr
educt
ion.Themi
nimum t
ime
el
ementi
ncol
umnsA,BandCi
s0,10and0r
espect
ivel
y.Subt
ractt
heseel
ement
s
f
rom al
lel
ement
sintheirr
espect
ivecol
umntogett
her
educedt
imemat
ri
x.
Tabl
e:Aft
ercol
umnreduct
ion

A B C
1 40 10 0
2 0 0 30
3 0 20 10
St
ep3:
Testf
oranopt
imal
assi
gnment
a.Dr
awt
hemi
nimum numberofhor
izont
aland/
orv
ert
icall
i
nesnecessar
ytocov
er
al
lzer
oti
mes( cost
s).
Tabl
e:Testofopt
imalassi
gnment

A B C
1 40 10 0
2 0 0 30
3 0 20 10

b.Countt
henumberofl
i
nes
I
fthenumberofl
i
nesi
sequalt
othenumberofr
ows/
col
umns,
theopt
imalsol
uti
oni
s
obt
ained.Thuspr
oceeddi
rect
lyt
ost
ep5.

St
ep5.
Makeanopt
imal
assi
gnment
Anopt
imalassi
gnmentshoul
dbemadet
ocel
l
swi
thazer
oent
ery
,mai
ntai
ningt
he
one-
to-
oner
equi
rement.
Tabl
e:opt
imalassi
gnment

A B C
1 40 10
2 3 30
3 1 0
0 10
0
0
0
Not
e:
Inopti
malassignment
,st
artwi
thr
ow/
col
umnhavi
ngonezer
oandcancelt
he
alt
ernat
ivezer
os(x)

Thepat
ter
nofassi
gnmentamongpr
ogr
ammer
sandpr
ogr
amswi
tht
hei
rrespect
ive
t
ime(
inmi
nut
e)i
sgi
venbel
ow:

106
 Pr
ogr
ammer Pr
ogr
am Ti
me(
inmi
nut
es)
1 C 80
2 B 90
3 A 110
Tot
alt
ime=280mi
nut
es

2.Adepar
tmenthasf
iveempl
oyeeswi
thf
ivej
obst
obeper
for
med.
Thet
ime(
in
hour
s)eachmanwi
l
ltaket
oper
for
m eachj
obi
sgi
veni
ntheef
fect
ivenessmat
ri
x.

Empl
oyees

I I
I I
II I
V V
A 10 5 13 15 16
B 3 9 18 13 6
Jobs

C 10 7 2 2 2

D 7 11 9 7 12

7 9 10 4 12
E

How shoul
dthej
obsbeal
l
ocat
ed,oneperempl
oyees,soast
omi
nimi
zet
het
otal
man-
hour
s?

Sol
uti
on:
Tabl
e:Af
terr
owr
educt
ion

I I
I I
II I
V V
-
5 A 5 0 8 10 11
-
3 B 0 6 15 10 3
-
2 C 8 5 0 0 0

-
7 D 0 5 2 0 5

3 5 6 0 8
-
4 E

Si
ncet
henumberofl
i
nesl
esst
hant
henumberofr
ows/
col
umns,
ani
mpr
ovementi
s
possi
ble.
St
ep4.I
mpr
ovet
hepr
esentoppor
tuni
tycostt
abl
e
Thi
sisdonebyt
hef
oll
owi
ngoper
ati
ons;
a.Sel hesmal
ectt lestent
ry(
element
)amongal
luncov
eredel
ement
sbyt
hel
ines
andsubt
racti
tfom al
r lent
ri
esi
ntheuncov
eredcel
ls.

107
b.Addt
hesamesmal
lestent
ryt
othosecel
lsi
nwhi
chl
inesi
nter
sect(
cel
lswi
tht
wo
l
inest
hem).
c.Cell
swithonel
inet
hrought
hem ar
eunchangedt
othei
mpr
ovedt
abl
e.

Tabl
e:Af
teri
mpr
ovement

I I
I I
II I
V V
A 7 0 8 12 11
B 0 4 13 10 1
C 10 5 0 2 0

D 0 2 0 0 2

3 3 4 0 6
E

Si
ncet
henumberofl
i
nesequal
stot
henumberofr
ows/
col
umns,
thesol
uti
oni
s
opt
imal
.
Tabl
e:Opt
imalassi
gnment
s

I I
I I
II I
V V

A 7 8 12 11

B 4
0 13 10 1

C 10 5 2

D 0
0 2 2

E 3 3 4 6

0 0
Thepat
ter
nofassi
gnment
samongj
obsandempl
oyeeswi
thr
espect
ivet
ime(
in
hour
s)i
sgi
venbel
ow:

Job Empl
oyees0 0 Ti
me(
inmi
nut
es)
A I
I 5
B I 3
C V 2
D I
II 0
9

E I
V 4

Tot
alt
ime=23Hour
s
108
3.Amanagerhaspr
epar
edt
hef
oll
owi
ngt
abl
e,whi
chshowst
hecost
sforv
ari
ous
combi
nat
ionsofj
ob-
machi
neassi
gnment
s:
Machine
(
Costi
n’000s)
)
A B C
1 20 15 31
Job 2 17 16 33
3 18 19 27
a.Whati
stheopt
imal
(mi
nimum-
cost
)assi
gnmentf
ort
hispr
obl
em?
b.Whati
sthet
otal
costf
ort
heopt
imum assi
gnment
?
Sol
uti
on:
Table:Af
terr
owr
educt
ion Tabl
e:Af
tercol
umn
r
educti
on

A B C A B C
-
15 1 5 0 16 1 5 0 7
-
16 2 1 0 17 2 1 0 8
-
18 3 0 1 9 3 0 1 0

Tabl
e:Af
teri
mpr
ovement

A B C
1 4 0 6
2 0 0 7
3 0 2 0

Tabl
e:Opt
imalAssi
gnment

A B C

1 4 6

2 7

0 0

109
3 2

0 0
 Job Machi
ne Cost
(i
n$)
1 B 15000
2 A 17000
3 C 27000
Tot
alopt
imalassi
gnment
=$59000

5.
6.SPECI
ALCASESI
NTHEASSI
GNMENTPROBLEMS
Cer
tai
nsi
tuat
ionscanar
isei
nwhi
cht
hemodeldev
iat
essl
i
ght
lyf
rom t
hatpr
evi
ousl
y
descr
ibed.Amongt
hosesi
tuat
ionsar
ethef
oll
owi
ng:

A.Mul
ti
pleOpt
imalSol
uti
ons
Whi
l
emaki
nganassi
gnmenti
nther
educedassi
gnmentmat
ri
x,i
tispossi
blet
ohav
e
t
woormor
eway
stost
ri
keof
fanumberofzer
os.Suchsi
tuat
ioni
ndi
cat
esmul
ti
ple
opt
imalsol
uti
onswi
t hesameopt
ht imalval
ueofobj
ect
ivef
unct
ion.I
nsuchcases
t
hemor
esui
tabl
esol
uti
onmaybeconsi
der
edbyt
hedeci
sion-
maker
.

I
nmul
ti
pleopt
imalsol
utons,nouni
i que0wi
llexi
statsomepoi
nt,r
esul
ti
ngi
nmor
e
t
hanonechoi
cef
orassi
gnmentandhence,
mor
ethanoneopt
imalsol
uti
on.I
tshoul
d
benot
edt
hatal
lopt
imal
sol
uti
onswi
l
lyi
eldt
hesamev
alueoft
heobj
ect
ivef
unct
ion.

Exampl
e:
1.Gi
vent
hisf
inal
assi
gnmentt
abl
e,i
dent
if
ytwoopt
imal
sol
uti
ons.
Machi
ne
(
Est
imat
edtimeinmi
nut
e)
1 2 3
A 4 0 0
Job B 0 3 2
C 1 0 0
Sol
uti
on
Thef
ir gnmentmustbeB-
stassi 1,becauseB-
1ist
heonl
y0t
hatappear
sinasi
ngl
e
r
ow orcol
umn.Hav
ing made t
hatassi
gnment
,ther
e ar
etwo choi
ces f
ort
he
r
emai
ningt
wor
ows,andt
wochoi
cesf
ort
her
emai
ningt
wocol
umns.Thi
sresul
tsi
n
t
wopossi
blesol
uti
ons,
asshown:
Machi
ne Machi
ne
(
Est
imat
edtimeinmi
nut
e) (
Est
imat
edtimeinmi
nut
e)
1 2 3 1 2 3
A 4 0 0 A 4 0 0
110
Job B 0
0 3 2 Job B 0 0 3 2
C 1 0 0 C 1 0 0
2.Thef
oremanofamachi
neshopwant
stodet
ermi
neami
nimum costmat
chi
ngf
or
oper
ator
s and machi
nes.The f
oreman has det
ermi
ned hour
lycostf
oroff
our
oper
ator
sfort
hef
ourmachi
nes,
asshowni
nthef
oll
owi
ngcostt
abl
e.
Machi
ne
(
Est
imat
edcosti
n$)
A B C D
1 70 80 75 64
or

2 55 52 58 54
at
Oper

3 58 56 64 68
4 62 60 67 70
Requi
red:
a.Det
ermi
net
hemi
nimum-
costassi
gnmentf
ort
hispr
obl
em
b.Whati
sthet
otal
costf
ort
heopt
imal
assi
gnment
?
c.I
sther
eanal
ter
nat
iveopt
imalassi
gnment
?Whati
sit
?Cal
cul
atet
het
otalcostf
or
t
heal
ter
nat
eopt
imal
assi
gnment
.

Sol
uti
on:
Tabl
e:Af
terr
owr
educt
ion Tabl
e:Af
tercol
umnr
educt
ion

A B C D
A B C D 1 4 16 5 0
D
1 6 16 11 0
D
2 1 0 0 2
2 3 0 6 2
3 0 0 2 12
3 2 0 8 12
4 0 0 1 10
4 2 0 7 10 Tabl
e:
Opt
imal Assi
gnment
s
A B C D
1 4 16 5 D
2 1 2
0
3
0 2
0 12
4
0 1 10
0
0 0

111
a.Opt
imalAssi
gnment b.
Oper
ator Machi
ne Cost
(i
n$)
4 A 62
3 B 56
2 C 58
1 D 64
Tot
alcost=$240

c.Yes!
Alt
ernat
iveopt
imalassi
gnment
Oper
ator Machi
ne Cost
(i
n$)
1 D 64
2 C 58
3 A 58
4 B 60
Tot
alcost
=$240

B.Maxi
mizat
ioncasei
nassi
gnmentpr
obl
ems
Ther
emayar
isesi
tuat
ionswhent
heassi
gnmentpr
obl
em cal
l
sformaxi
mizat
ionof
pr
ofi
t,r
evenue,et
c as t
he obj
ect
ivef
unct
ion.Such pr
obl
em maybe sol
ved by
conv
ert
ing t
he gi
ven maxi
mizat
ion pr
obl
em i
nto a mi
nimi
zat
ion pr
obl
em byt
he
f
oll
owi
ngpr
ocedur
e:
i
.Findthelargestpr
ofi
tcoef
fi
cientintheent
ir
e.
i
i.Subt
racteachentr
yintheori
ginaltabl
efr
om thel
argestpr
ofi
tcoef
fi
cient
.
Thet
ransf
ormedassi
gnmentpr
obl
em soobt
ainedcanbesol
vedbyusi
ngt
he
Hungar
ianmet
hod.

Exampl
e
1. A companyhasf
ourt
err
it
ori
esopen,andf
oursal
esmenav
ail
abl
eforan
assi
gnment
.Thet
err
it
ori
esar
enotequal
l
yri
chi
nthei
rsal
espot
ent
ial
.Basedont
he
pastper
for
mance,t
hef
oll
owi
ngt
abl
eshowst
heannualsal
es(
in$)t
hatcanbe
gener
atedbyeachsal
esmani
neacht
err
it
ory
.Fi
ndt
heopt
imalassi
gnmentandt
he
maxi
mum expect
edt
otal
sal
es.

Ter
ri
tor
y
I I
I I
II I
V
A 42 35 28 21
esmen
Sal

112
 B 30 25 20 15
C 30 25 20 15
D 24 20 16 12

Sol
uti
on:
Conv
ertmaxi
mizat
ionpr
obl
em i
ntomi
nimi
zat
ionpr
obl
em bysubt
ract
ingal
lel
ement
s
f
rom t
hehi
ghestel
ement(
i.
e42)

Thus,
theequi
val
entcostt
abl
eis:
I I
I I
II I
V I I
I I
II I
V
A 0 7 14 21 A 0 3 6 9

B 12 17 22 27 B 0 1 2 3

C 12 17 22 27 C 0 1 2 3

D 18 22 26 30 D 0 0 0 0

Thus,
aft
eri
mpr
ovementoft
het
abl
e,t
heopt
imal
assi
gnmenti
s:
I I
I I
II I
V
A 0 2 4 7

B 0 0 0 1

C 0 0 0 1

D 2 1 0 0

Thepat
ter
noft
woal
ter
nat
iveopt
imalassi
gnment
samongt
err
it
ori
esandsal
esmen
wi
thr
espect
ivesal
eisgi
venbel
ow:

Assi
gnmentsetI
Assi
gnmentsetI
I
_
___
___
___
___
___
___
___
___
___
___
___
_
_
___
___
___
___
___
___
___
___
___
___
__

Sal
esman Ter
ri
tor
y Sal
es(
$) Sal
esman Ter
ri
tor
y Sal
es(
$)
A I 42 A I 42
B I
II 20 B I
I 25
C I
I 25 C I
II 20
D I
V 12 D I
V 12
Tot
al=$99 Tot
al=$99
_
___
___
___
___
___
___
___
___
___
___
___
_ _
___
___
___
___
___
___
___
___
___
___
__
Exer
cise
Fi
vesal
esmenar
etobeassi
gnedt
ofi
vet
err
it
ori
es.Basedont
hepastper
for
mance,
t
hef
oll
owi
ngt
abl
eshowst
heannualsal
esmani
neacht
err
it
ory
.Fi
ndt
heopt
ional

113
 assi
gnment
.
T1 T2 T3 T4 T5

S1 26 14 10 12 9
S2 31 27 30 14 16
S3 15 18 16 25 30
S4 17 12 21 30 25
S5 20 19 25 16 10

C.Unbal
ancedAssi
gnmentpr
obl
em
TheHungar
ianmet
hodofassi
gnmentr
equi
rest
hatt
henumberofcol
umnsandr
ows
i
ntheassi
gnmentmat
ri
xbeequal
.Howev
er,whent
hegi
vencostmat
ri
xisnota
squar
emat
ri
x,t
heassi
gnmentpr
obl
em i edanunbal
scal
l ancedpr
obl
em.I
nsuch
casesadummyr
ow(
s)orcol
umn(
s)ar
eaddedi
nthemat
ri
x(wi
thzer
osast
hecost
el
ement
s)t
omakei
tasquar
emat
ri
x.Af
termaki
ngt
hegi
vencostmat
ri
xasquar
e
mat
ri
x,t
heHungar
ianmet
hodmaybeusedt
osol
vet
hepr
obl
em.

Exampl e
MEGA pr int
ingpr ess,apubl i
sherheadquar teredi nAddisAbaba,want stoassi gn
threerecentl
yhi redcol l
egegr aduates,Mar ta,BakchaandHi ruttoregionalsales
distr
ict
sinMekel l
e,Bahi rDare,andDi reDawa.Butt hef i
rm al
sohasanopeni ngin
Gambel aandwoul dsendoneoft het hreetherei fitweremor eeconomicalthana
mov etoMekel le,BahirDarandDi eDawa.I
r twi l
lcostBr.1,000torel
ocateMar tato
Gambel a,Br.800t orelocateBaklchat here,andBr .1,500tomov eHir
ut.Whati sthe
opt
imal
assi
gnmentofper
sonnel
toof
fi
ces?
Of
fi
ce
Mekel
le Bahi
rDar
e Di
reDawa
Hi
re
Mar
ta Br
.800 Br1,
100 Br1,
200
Bekcha Br
.500 Br1,
600 Br1,
300
Hi
rut Br
.500 Br1,
000 Br2,
300

Sol
uti
on
Tobal
ancet
hepr
obl
em,
weaddadummyr
ow(
per
son)wi
thazer
orel
ocat
ioncostt
o
eachci
ty.
Ci
ty
C1 C2 C3 C4(
Gambel
a
)
P1 800 1,
100 1,
200 1,
000

Per
son P2 500 1,
600 1,
300 800
P3 500 1,
000 2,
300 1,
500
Dummy 0 0 0 0

Tabl
e:Af
terr
owr educt
ion Tabl
e:Opt
imalAssi
gnment
C1 C2 C3 C4 C1 C2 C3 C4

P1 0 300 400 200 P1 100 0 10

10
14 0
P2 0 1,
100 800 300 P2 0 700 400 0
P3 0 500 1800 1000 P3 0 100 1400 700
Dummy 0 0 0 0 Dummy 400 0 0 100
Thus,
anopt
imal
assi
gnmentcanbemadeatzer
ocel
l
s(squar
es)
.
Per
son Ci
ty
Dummy(
Noper
son) Di
reDawa
Hi
rut Mekel
le
Bekcha Gambel
a
Mar
ta Bahi
rDar
e

Cost=Br
.(
0+500+800+1, =Br
100) .2,
400

D.Rest
ri
cti
onsonAssi
gnment
s
I
ncer
tai
ninst
ances,i
tmayhappent
hatapar
ti
cul
armat
chorpai
ri
ngmaybeei
ther
undesi
rabl
eorot
her
wiseunaccept
abl
e.Forexampl
e,anempl
oyeemaynothav
ethe
ski
l
lsnecessar
ytoper
for
m apar
ti
cul
arj
oboramachi
nemaynotbeequi
ppedt
o
handl
eapar
ti
cul
aroper
ati
on.I
nsuchcases,t
hecostofper
for
mingt
hatpar
ti
cul
ar
act
ivi
tybyapar
ti
cul
arr
esour
cei
sconsi
der
edt
obev
eryl
arge(
wri
tt
enasM or )so
ast
opr
ohi
bitt
heent
ryoft
hispai
rofempl
oyee-
jobi
ntot
hef
inalsol
uti
on.Whensuch
ar
est
ri
cti
oni
spr
esent
,al
ett
er(
M)i
sof
tenpl
acedi
nthet
abl
eint
heposi
ti
ont
hat
woul
drepr
esentapar
ing.Anal
ysi
sisper
for
medasusualexceptt
heM i
signor
ed
t
hroughoutt
heanal
ysi
s.Thati
s,M i
snotusedi
nanyr
educt
ions,nori
sanyv
alue
addedt
oitorsubt
ract
edf
rom i
tdur
ingt
hecour
seoft
heanal
ysi
s.

Exampl
e
1.I
nthemodi
fi
cat
ionofapl
antl
ayoutofaf
act
oryf nesM1,
ournewmachi M2,M3and
M4aretobeinst
all
edinamachineshop.Ther
ear
ef i
vev acesA,
acantpl B,C,
Dand
Eavail
abl
e.Becauseofl
imit
edspace,machineM2 cannotbepl acedatCandM3
acedatA.t
cannotbepl hecostofplaci
ngofmachi neatplacei(in$)i
sshown
bel
ow.
Locat
ion
A B C D E
M1 9 11 15 10 11

Machi
ne M2 12 9 - 10 9

M3 - 11 14 11 7

M4 14 8 12 7 8

Fi
ndt
heopt
imal
assi
gnmentschedul
e.
Sol
uti
on:
Ast
hecostmat
ri
xisnotbal
anced,addonedummyr
ow (
machi
ne)wi
thazer
o
costel
ementi
nthatr
ow.Al
soassi
gnahi
ghcost
, edbyM,
denot tot
hepai
r(M2,
C)
M3,
and( A).

115
Appl
ytheHungar
ianmet
hodt
osol
vet
hepr
obl
em

A B C D E
M1 9 11 15 10 11

M2 12 9 M 10 9

M3 M 11 14 11 7

M4 14 8 12 7 8

M5 0 0 0 0 0
Tabl
e:Opt
imal
assi
gnment

A B C D E

M1 0 2 6 1 2

3 M 1
M2 0 0

M3 M 4 7 4
0
M4 7 1 5
0
1

M5
0 0 0 0 0

Thet
otal
mini
mum cost(
$)andopt
imal
assi
gnment
smadear
easf
oll
ows:

Locat
ion Cost
s($)
Machi
ne
M1 A 9
M2 B 9
M3 E 7
M4 D 7
M5 (
Dummy) C 0
Tot
al=$32

116
Exer
cise:
1.Acarr
ent
alcompanyhasonecarateachoff
iv sa,
edepot b,c,
dande.Acust
omer
i
neachoft
hef
iv ownsA,B,C,D and E r
et equi
resacar
.Thedi
stancei
n(i
n
ki
l
omet
ers)bet
weent
hedepot
sandt
ownswher
ethecust
omer
sar
e,i
sgi
veni
nthe
f
oll
owi
ngdi
stancemat
ri
x:
Depot
s

a b c d e
A 160 130 175 190 200
B 135 120 130 160 175
Towns

C 140 110 155 170 185

D 50 50 90 80 110
E 55 35 70 80 105

How shoul
dthecar
sbeassi
gned t
othecust
omer
sso ast
o mi
nimi
zet
he
di
stancetr
avel
ed?
Answer:
A e
B c
C b
D d

Mi
nimum di
stance=570km

2.Anai
rl
inecompanyhasdr
awnupanew f
li
ghtschedul
einv
olv
ingf
ivef
li
ght
s.To
assi
sti
nal
l
ocat
ingf
ivepi
l
otst
othef
li
ght
s,i
thasaskedt
hem t
ost
atet
hei
r
pr
efer
encescor
esbygi
vi
ngeachf
li
ghtanumberoutof10.
Thehi
ghert
henumber,
t
hegr
eat
eri
sthepr
efer
ence.Cer
tai
noft
hesef
li
ght
sar
eunsui
tedt
osomepi
l
ots
owi
ngt
odomest
icr
easons.Thesehav ebeenmarkedwi
thaX
Fl
ightnumber
a b c d e
A 8 2 X 5 4
B 10 9 2 8 4
lot

C 5 4 9 6 X
Pi

D 3 6 2 8 7
E 5 6 10 4 3
What shoul
d be t
he al
l
ocat
ion of
t
hepi
l
otst
ofl
i
ght
sinor
dert
omeetasmanyper
for
mancesaspossi
ble?
Hi
( nt
:Thepr
obl
em i
stomaxi
mizet
het
otalpr
efer
encescor
e)

117
Pi
lot Fl
ightNo Pr
efer
enceScor
e
A 1 8
B 2 9
C 4 6
D 5 7
E 3 10
Tot
al40

118