3-0

Module- 3

Cved Method benen al method, con change problem, napsock

Pro blem, Job sequeneirg wih oeodkined a- d, h .3, 4S
Manimm cost spanning breas: prim's algosthm, kruskal's algoithm
,9.a
Singe sounce shovte paths Dijkstra's algovithm I93
Omal te problem Halfnan Trecs and coolo 94
and Heaps ovt 1- &.4
ans bvm and ConGue pproach Heaps

- StStraightfovumd desgntechrique
Greod rmethod
Genenal nedhod
o oblan a subset that satke
Fov h 'puts, Yesuire l

Some constraints
the gn»tvasnt calleol a
&ubse thot saBske
Bny
eable solus'on that eithe maximize ov

Find a
eoible 80lukon
s0lukon dees b caltd
unmiz gn. a okjeciue fn. A eas:ble
Opkmal solukon.
hat one Can deuise an
The heely methoel suggeats
that wore in Stages, consiolering One I/p at
a
goithm
'ne
A each staçe a dectsion ?s made Yegardtrg whethe, a
solukon. Done by consdering the
PavHcula I Ip is in an
opmal
n Ovden. datemined by Some selaetien prOcedure
Ip s an

If Ehe inclusfon t h e nexB 7)p inu bhe parkal

Yetult tn a n inleam'be s 0 luhons
Consbructee opkmal Bolukon oi
Ehen th I}p is not adoed o he pavhal salukan.

Othevwte it is acloed
 SelecHon proceduno baso on A0me opkm3aon meastge

Meooluu may be he objecbue fnckon

Sevenl dfforent opHm/zakon meau may be plausibe

r a an poblemMost 2 bexe howeven wil res ult in algondhnm

Bat qenevale Aubopimal solukon

Thi vevston of the 9heesy techn'que * alleol

Subset poadism

Conb abshackon o dhe Bubset pavaoigmn

AR
elgost
Greeoly l a , n ) /atn) pats
afi 'n) centaud
contaudho n

Irikale dhe Aolachkon

fov z e E I E 9 i:=Ib n do

poBelect
select (a), Sebeb an p kom al) arrereest
f Seastble (Boluhorn x) dhar aastbe fs a bolev alrd
Soluhon= union (80luken ) ; Tunetitn dhat dotems

3 Whatla an incloeled inbo

Tebn 0lukon dhe soluhßon vecha
and
3 urndn onbines x with dre solulson
pdolo the ohyekp henekan
Fov pmbero that do not call Jor e
elechon e,, an
selkchon an

Opb'malsubset in dhe greeoly methoo, mate

oecion hg coroiolerng
e p S tn Aome d
ovoler
Each olectston s maoe ing an
opimizakon Cieion
tha can be Compubed usng olecssfos alreaoly maole This
called as
rady rmeod,ovoleirA panadiGm
Bubset pa-adt m -

Cnapsacl pmblem, Tob

sequenchq wi deaolire,
minimu m Cost Apanning ree, Contai' ne. oadiog
(Yee Spliti
ve tes
0Yoloira paaoligm -

Sngl Source shoatest padtho, opkmal 8 oeg tpo, on

opima mO pattecs
Coin chnge pme 3-
A chld bnys Canely Valued ot los han D 100 arcd

gve a 2sl0o bill b dhe cashien

Caxhte wtha b vehn charge unirg he ouvent

sSune thak an untmiked Aupply 86 rpes,

10 vupee» and 5 Yupees Cno cofns i's avaflable.

Cashre Conatruct the change in sages. In each

bo dhe change Ps selectrd wg
Stace cofn s adolesl
hecoly citerton
incveae the ptas amounl Char
At each stoce
much ao pssHe.
Conwucted by
(ie. he chongeven exactiy
To asuru ysasibilih
dte solukan, bhe selackeel cofn
equals he destved amount ),
qfuen so fan
the totas amount Charg ue
7 chong
caue
Srould ro
deireo amOunt
exceeo the nal
n 's du the chu.
Ehat 54 Nupees change
Supros
Suprose
seleokrd are do Npee
The rsk a
the 3vo
3ro

A do Pupees cannot be selecked ov

selechon
resul&nan ineasible seachon
berase Such
betase a

(Change
eicaæos STrupees)
upe
The Ehard Pank selockeol is a 10 Rupees, bhen a

5 Pupee is selecked ard nally 3 One nypee coins are

adeleo » he chang.
 The ghccody metiood has inhutkve ppeas in dhat whj),
he inhuihon
Conahnuct he change hy wing a Rbategm hol
d w dhe
tello us shodd reult dhe ewent (ov at leart close
euvert ro Rupees ov Cot behg 9ven Out
Ove dhal Hre ecoly algodtn olehbeon, inolao great
Chan cerd dhe tewet nam e
Knapsack ppbl em ) cot

iven n objects ano a

e nap Sack ov
bag.
obecthas a aweight Cwr and tbe knapSaclc has a

Capath m
IP nackon:, Og x:<I
a
objectiPs placeed
in Be
krapsack, bhen e prokt Pra eon
s

Objechve s g a knapsack bhak maximzes bhe

tobal pmßt earned
krap sack capachy m. tobal wegh g all chaten
obyect o he atmest m.
PmRt arol wegh au posihve umbe

opmzr max inmiz á P 0

Conobrant subject bo 2 wX; m

and o x:4), 5:sn

FeaibleSoluhon iS ary set (X|-.. An) Sakstng and

Opmal oluhon is a easble 8oluhon 1oY whrch 0

moxtmized
 weight dhe obyecks chosen Bhould not exceed bhe
capociy knapsack
objec ae ahanges in oleerwasta 0volon prJw

COnsidlor bhe \lo cotng fratance 9 the enapsacl pmblem

m a0

(P:Pa Pa) l3s 3k s)

(to Wa wg) ( 18 S 10)
P 2as IS PL 2PlH)
wi 8 1S 10

1 39 S

S 39
P 1S as
w
16 18

X=I oY YC TC C - w r X
ObectC) wt Pi
*Pi
a0-IS1 = S
IS

S- lo O S= o
a S 0S O.SXIS7.5

3 18 as

Frachon (, a *z) =
(10S o)
(0 0s)
Prot ah+ 7-s 31 S
.obtain dhe Ophmal 8olukon br he
krapsack prubem us
Prg
honsluy metood

P Pa Pa 2
(30 o I23
3s) i 18 IS 10
P 30 Pr 30 2r 18
w
as

6 3S M1S
?i lo S Is 7 b 18 3
3S 4o 30
w lo as

Obgect t wi P = | or YC PRE
wi pi
3s 3S 4o-/0 30
2 as
30-aS =S
3 30
S03S
a
S S-S 0

rachono (X, a a3) -(Io3s)

(0-as 1)
P BE =
3S 4h0 +7.s =
83-5
The Bum
Bum all the we'ghts is Sm Ben Xf =/, 18isn
S Gn opkmau 8olubon AssSume e Sm weighb exceess m.
NO all Be X:'s Connot be .
PP opknmas olukons cwfIl bhe knapsoc exaetly brus
becawe aluns i'ncveae e conbskukon some
eg obfecti
a nackona amoun unsl the tas weight s exactly m
The meted h 6lkau'n dl Boluon steeol
used reoly metal
a

calee ar each Abep ( e t posTbly he atone), Choose h toolua het obes

Vdu de mea
nvone dhe doseebe .
Wk wewd
 Vod grnady Jemptack ( Jloak m , int n) MMa lo
oi 4 } 53
Pi 4o 4a as 1a
fo int ia1; i Len; iw)
ui = 5 o 40 30

float Yc=m Pi 3o 2o loo % 5o

b(i1i Cen) it+)

wL:] >rc) object Cannot be &electrd
eak
iSeleet bhe object
CYC -wli), \W d rematnirg capoig genapscek
3
f(iL=n)
Li)- *<lwt:) \ Jjrd yaachon )object selucteol

Sob Sequenc Png w öh

oleadlineri
tven a
3e n
jobs
As sociaked widh job is an integen olesdlire drz20
and a
pokE pi >o
Fov anyjob ', he prmkt Pr i eavred arol o f

te Fob 1s
Completeod hy r oleodine
To Compete a job Ore has h pmes &e job on a
mlc hv One unit Bme
Only one machine s ava'lake br proces jos

A
at ble 8oluton hv dhe pmbem a ubset T 3
os Buch that earh job in tho Subset can be
COrpletod by
deadlhe.
The Value enrble solukon I dhe &um te

PoRts dte johs n 3 oP

 Pn
ophmal solukon Ps oibe solukan widh ma volue.
Poble m huolue dhe iolenbkrakon Auhset k& he
g a ,
tt
Saset paaoigm

Solve the
hllow ob c_uencing problem (mau'mizms
hllowna the
poktg
POK
Complekng johs bejove b-etY oledlines) ustng qraeol
alonthm
N Nojoks= H

Prolts asoctatrol uit4 jgbs CPP2a Pa) l1oo 10, Sia9)

Deedlines asoctated with jpbs (didadz
d4) = (2. h 2,)
rrans pmlE p; in dencendira
Oelon
ProRt 00

P PA Pa
deadlene
CYeole an
aau i c h Abre,
the jobs
3

Add h job in avay at bhe index olenoted by deosle.

H job p, The oleodlne o dhi soh n a. Herce inet P in
avey at ara index .

P
TXE job p Insevt i t al inax

Pa P
Next job s
Pa wih dleodlne a. But amay i3
alroas
Occupred Hece discand P3
 Solukon an pnceeo
39
3 s opimal Only jobl anol

ar he value is 1a7.
Thee jobs mu be proceseod tn he Ovole job k pollaweod

b Job The begin a Bne 3eo an that

procesina 2 b
jobl is
compleked ot Bme
a.
The eatHe soluken and ther values au

value
eaile solukon pvaces4ueee
(1a) ,

C3) 3 oY 9,)
a pmdso lasn
1,4)
(2.3) 2,3 aS
3.4) 43 42
(1) 100

(a
(3 S
4)

lo obtain an 0pmal Bolukon. ovmulate an

opimgakon measu to deke min how bhe next joh rs chosn.

Choose the objeckue Funchon S P: a the opmizabon

measuu
inq dhi meamure the nex job o incluole?s he One
Ehat
incveae»2 Pi he most Subject h be Corsbraht hat

Che
reulkrg T 's a easble solubson.

Constole
Onstlenjobs t non incveanfng Oolar c dhe
e Pis
pis
Hgh Aovel decypkom job sequenctng algovid
eody ph ( int dl), set I, int n)
3'seAt q, jphs l het con le amp lebe by
dhaleodlina
or (in i
- 2; i<on; itt)

all johs in
Jus*} can be
completrd by thetr oleodlire)
33
The Gbove algoidhm Corshnucts an
epkma Ael 2
jobs tha can be proces sed by dheir due ime
In job Ehere
se_uencng a 2 poKsTble parame ters
in term cwhrch i
comdexty Can be measwned
wse n
-no johs
S-
ro jobs incluoleol in dhe siluson J

lgpvidhm ov seguercing unit ne job toidh oleodbns and pmkk

int 3s (int dti, i'ntjt3., int n)

dlo 3LoJ=o, alz

Wincluoe job)
i'nt e=l
for (int i -a; jtenj
itt)
\constolen jobs ih
nonincreosing Dvoler pli). Fro pos thon k
and chek
easiblhhg insebon .
Hah Aovet decyphun 2 job seguenctng alypridhrn

eody ph / int d), set J, int n)

/3 s e Ael 4, pbs lhol con e Camplebi py
dhe' oloodltma,
oY (inti =a; ion itt)

all johs in
JuSi} can be comperd hy thetY oleodire)

33
The above algoidhm conshucts an
opkmal Ael 3 2
obs thab can be pvDces sed by dheir due ime

In
n job sequencing there au 2 possible parame ters

in bevm e which i Compexig Can be measuned

ue n-no j o h s

no jobs ineluoleol in dhe seluhon J

Hlgovsdm ov seguencing lanit me job cordt oleodbhe anod pmki

int Is (in di1, intjl), int n)

dlo-3lo)=o , ni Hali'z
incuole jo
int e=l'
fov (int i =2; i
tenit+)
constolen jabs
non'ncveoning DYoler pli). Finol posikon bri
and cheek eoibikhy cinseaon.
in Y-k,
 ohle
((d[1) >d(n) (a[í))
Y--

fnse. fnto 1 |
l» () slop 1 d

Yen (kD)

he seguence 3 w be insetrd if ond orb, f

3EIJ! =t. Means thas he Tob 3 will be procesed f ?E t: in
Lwithin dha deod len
The Compukng kma taen hy dhis alg OlrP) beroue
ohe baot opeakon Compung equere n a y it wttn

2 nokre to Jeops

AEE
hS ,
(P. a Ps Ph Ps) =
(20,1S, lo. S, D and
l(d. op ds d, ds)
*

(2, 3,1,3,3)
3
P ao Slo S
di 3 33
 Select 1=1, P; >O , di- a

Sobl twd pmkt ao

Selecli,P:Is,d
Job I and a widh pokE a0 4S =3

i 3, Pi= 10, dr=i Th iob is

nol Beectrd,prokt won't
Selke
.

chongt
Selr i H Prs ,diz 3
34

with pm kt o tiS+S = 40
Tobs I and

Rest h e jobs ane discancool as he deodlhes ar

Oven
OpHrnal &olukon 5.3
otal prRt =
20 4 IS + S=o

dbsned slets ob coraicdernd chon pro&

a)
To,Da) tanro t Rt, geet
3
.33 Lo3)a) 35

1,3.43 [oS.2Ja,3) rejeot

he ophmas 8oluktr is 3-S1,3,A} h a prkt *o.
 3-
Minimum cost8panning r e

Pim'S olaovdhm
atven n
points, Connect bhem tn be cheapest poseible wo
So hat bhere be padh. bet
coln a
evey paiy points
Poins- Veitre aph 3a
Possidle Con neehon-gnaph's eolga
Connechon Cosk eolg uwezhts-

Spanning breet
pahnig rea a Conhecteol gaph is Connected acycbe
Saanaph that conains all bhe vevkce be ghaph
A n a a qsaph Ps a Skgaph ih-ch ?s basiay a hree
Minf C n:n Vebre, s, a containis he ctevsk.
MST
MST CG weightrd connecked Gsaph s i& Spanning ree

b Dhe Smalk ueght whew e ue'sht a bee is d e n e l a s ue

Sum
bhethe ueht on all i6 esge
MST
po bem Ps the
pmblem 2 1indins a minimum
Sponning bre bv a gn weisheo connet trd aph

3
 In eha totve ohstacles
Beav ch appmath ,
nospannig kes qows oxponer'alid he gaph Aiz

(at leant fov dlonse ph)

bres Bv * bo
genenakng all Spannng a
g eph
In Pim's algosdhm,

Constyuct a nin ST broush a sesuence exparoins Subbreas

The inikal Subhe in Such a

Se_uence Cons's a singla
Vevbe electrd avbibvansly Prorn be set U e rap}s Veyie
Alsoibhm stops after all he 4aphs vehce haue been

ncluoled tn be es being Cororucted,

bex on
each ih
hy eraetly
vev
Hlg exparob a ke One

íteakoe
Otak n0 4 tteakon is n-I when n is dha no veyhres

Bthafph
The tre Gennotrd dhe als is obtained as he et
ep wred the kee expanston
ghaph will have ony one minimm spanning hee f ane

Ondes f &he
the weshk asociatel torth all the
eolke in he qnaph ane

dsknct
In cane d HST doe not malu complete Sense.

Pppi cobions

helpkl in 0nin9 apphcabom

'sl alunyp jdd a ms7

ST ued ceznns etkce -|* aprd,

nehepr olokg
 ind he m nimum Syying ee v dha
tblowing ghaph
9vng he hactrg
6 | 3 AS
o
o o

(a So

So &'o
So
3 4lo o O 30

So do30

Finel he eas cost

ov edge (o.2) = 1o

Consden Source =o oY 2

Take Souce = 0

dlo o O 6 O
dl 6o60 6o 4 A PL3=o O
dla) 1o LD PLa)= o O

d33- PL3) o O
o 30 30 30 PL43 o O S
dlsj So 50 SD SD PLS)=o

S V- S
dlvJ= min( dtU), costlustva) P£vJ= u udlul upluo
I,2.3, 4.S
0,a 3,4.S dli3 - 6o
dl3)=% PLSJ 2 S.So Sa

dls) So
0.3,5 ,34 dl 6o Pl3)-S A30 AS
d[3) Po PLA)-S
dL4 30
O,2. S, 3 dt zho PL-4

O.2.S,4.) 3 dl3) ao Pi3)=I 3 3,

0,2,SH3
roeec, hy sdykrs
Coa h d be daleen &r na n s Gru
Spanningbe bhained by(,0) (S,a) (4is) (1,4) (3,1)

Cost bhe MST =150

3
So
3o

Iree Vevkce»
Remaining vevkces
al--) bCa 3) Cl-,6) d(-,) elai6) fla.S

bla3) clb)d(-1) ela6) P(b4)

clb.) dlc.6) e (a.6) f(b,4)

fCbin) d(fis) e Lfa)

elfia) d (f.s)

dCs)
nemes ma vete

Tobal cal =iS

0oblafn an eolge t»hh lennt o Be cost
adotarg moih
min 99,
Soue-0,
fov it-t hn-do (
o io n-id
(afi) o b» allj) smin)
min alil)
Souvefi
endl if
Cnod v
Cnd v

JniHalzahon o rd min: &T\ ('o

ov i-o o n-1 do

dtil-wl souncr, ) (
Pl)Bouvce
Fird msT
S[80ue)
Sum o e)

Wrd u ond dlu) such dha! dlu) fs min cro ue V-s

aod u h S

Scec an
eolg wrdth leak cost
Add cost selscted edgs to get botal cost e MST
1eY ewenn ueV-s de
if wlu, )<dl))

dlv) luw)
PtEu
erd if erd ir erd
hor
 lea»t cast nom bhe cost askjar
3-
O Obtain an eolge otth
mab

min99
Source -0,
tov i-o lo n-I do
to j- o b n-i do

min ali)Lj)
Souve fe i
S
endif
Cno bv

S
TntHalzahon o 1nd min 87
fov i-p h n-i do
St:) eo
dliewl80wnce,
PiDEBource 3

Firod msT
S[80uce)
Sum 0 (13
or ie o n d

r d u and dlu) such dhat dlu) fs min arol ue V-s

aold u b s
Seect an eol wrtn leamk cest

Add cost selecked edss o get boal cost HST

0Y even e V-S edo

if wLu, v) d[))
dtvewluw)
PEEu
erod f erd hv erod tov
 S (8um :44)
whbe sponning ha d not ex
clhe
he erists and p r he MST
Cwste spannig

ime Coplexly pio's algowhm, (n®)

Hlg.
Olval)
Apends mest d re in sekcky dhe edye wrh min lengdh- Hoa he or

Kruskas's alprithm
Alwa yels an opimal olukon
MST or a wethtd Connecbed qeaph h(V,E) as an

e Sabtah wm lvj-I eelge B7 whrh bhe Sum the eo wezh

is the smalet
Algorithm beginoby sakig dhe *ap's eolgos in non olecoa,ihg
OYolen e bher wegh
Stasbirg wib dhe erphy Subgsaph, it scan hi sovteol s t
aeldirg the next eoke on the s t b be curret Bnbghapk # Bach
an incluson doe not cveale a Cyce aneSimpy sppin he

edee ohevcose.
Each Connecr Comporen a Bubaph gentabedl hy
u ta' s alg ts a kee bco t hao ho
cycb-
Csua's ala han chel whe tme bold'on 4 he neet edje h e
create
olneodly hpked would a cu

Alwags the min coul esdge has h

electrd: Bus t'snet heceAmn
dhot oleckes bpmm ool is nonant
 3-19
*Tve edge Sovbed st colgos
be
be ef ab bf cf af of ae cd de

3 S

bc

3
ab3

dfs

Krseas's alg. is ohe 7,a ho appliahonhat reguire a

dynomc Parhhon Somen el2men Set s ino a colle chon e

drsjotn Subseh S S. Sk

Disjoint Subsek ard Cunion

nd algosdhms
malceset (x- c veate a One eenment ee x}. lhi opevakbn con be

Pplied bo each of the elamenk set s

On oxe.

ind (x) - Yeh»m a subset containg

un onx, -

constvuct the urion of the digjont Subsets Sa arc

s
OntainigX ac y Yepechet4 and adlls it to the colleckon

to veploa Sx andSy twhich ae olelted rom i

Exemple,
S ,23,h.S,6 S

S
malke Set : ) = S3a3 33 S43 2r} 163
uan(1,). uron ( S.a ) 5 ,4) s2) $3313
=

union ( S) ardd 6
urion (3.6)» 5 i,4s2} S363
S4,8,63
O
Find the MST ing kukal's algosdhim

16

Eda
Eola cost dektrd tndlo) olp union(ij )
edg J-hdcv) (u,)
(u)
o 2 3 4S 8
6-3 6,T
67 O a 3 A 5 668
8 3.8 8 O 2 3 4s 6
S S6 S6 sI6
-1
O 2 3 s s 62
o O 3 S S 6
2-5 8
S O2 3 S
2.S
6-8 1 6,8 ,
discad O 3 a S
a-3 2,3 ,3 a3
4-8 3.8 , d'sad a4 aS6
0-4 1 6,7 b.4
-a 16
3-4 18 d'sad 4 S
3 6 ,4 34
4-S
- 3
3-S

eolge (6.7) Btore s S7

tov
edaa ( 6.8) nd fov eolge (6,8) s (213), 2 and 2
are sam
eo ( 618) muat be discanded, i4 it s Selected, ves uet in a Creeit

18

16
Cosk ST = J4
 3-
Algovdhm
Count-o nihal no edge,
keo poin fvct selecledl edg T
Sum-0
Inihal cost HST
br id v n elo CYrate a
wre,t wrh h
erhte
Ponentli) ei
Fnd hr

wkle ( (
Coumt n-) and (E+¢))
Seeet an
edge (u) widh least cosl
ird (uipanenl ) ind be oot ur he veler u
j inl v. pavent )
f roob au d'tleent select the edge lu iu)
as he eo MST
)

Count ++
Iup date no edge selecked fov HST
Sum-Bum+tostlu,) pdoe be cost MS7
Union (iipavent ) )
end tf menge a he wh rot
iand j
Pnd wlse

count! na)
Spanning ha does hok extsk
Yen
end
Spanning he 1s 6
toY (ie obn-2)do
oite (iLo), tfrya) 3
end hr
whle cost e
spanning hee, Aunm)
 Phro's alsovidhm
Tov obtainng Hst by SelocHng dhe adjacnte v ho

alraasly Seloctnol veykre. prinms widh nodes

algorithm inihialzps
kTuka's algovdhm
bY obtaintg MST but t ?s not necessany D Choose adjaeont
Vese aluady Belecha Vevkce. kruskal's algoidhm iniBalzes oih eolgo .

Time Complexihi

If we have E edee, wilN take (E Los E) lime hw Aov

Ehese edce in non
-der@asing Uvolen
In dis pin ets, there au n opevaionm on dis^o int selE
iLake O(n) Eme
So vest 2
operakon woulol in wose case o(E) hm e

So o(E Cos E4 E)

Source Bhovles padhs

Dtkstva's algorthon
fov a
roY a gn Vevbex calle the s0uvce ina werghted

mected aph, nod hovtest palh» to all ohen verk'ces

The alg.s oppkeabe h eph coth

hon negahe
weighh onla

verkce in ovole
Difisha's a
al inol the shrtest patha bo a aph's
ndhetr osstarce nom a n: 80mee
tvs irols bhe Bhovtest pah om bhe sourre
h a vevtex. nearo bo ?, then ba Secorol neanesl and So on.
Apd mS euk in drining a ow cost Coryate nehworle

hee Ms ndin minimna distanee be a hers

Solve bhe follainq Sirglo Aowhee shov tesl palh pro blern onu mra 3-19)
d
.Ves tex S a the Sance

20
o 15 1 4S
do 35
3 10
30
ao
S
Stat w Hh sowre nooo S

Jn' al t e nce
dlo) ob 49
dL o
dla) o
29
dla)a4
39 34 3H 34
d Ls O O

S V-v-S d[-mirn (a[v), 4l») +o[a)tv) nd u.dl) Aueh

hak dlu)
mi'nimum ard v EV-S

5 0, 1.2.3,4 3,
5,3 0,1,2,4 dlo= mir ( o, hto) o0
dj 4 , 14
dla) w
d[4)39
5.3,1 D,2, 4 dle)- 2, 39
dLo) 29

du) 3
5,3,1,3 o, dlo) 49 3
dlA) 34

S3,1.24 d o)= 49

5,3,1,2,0