Shovlebt dtskane om S » all ohes norle,

dlo) 49
- d)

d)*
d[) 3y
d[s)o S.
8bYPd n the any To
e vevHe that ve vis?rd a

locahon in ic inihalrd 6
tont w Yh ro VeyHen ane vised ad each

Sto

Algovidhn
dlie-cost LApure,t
SC-o

tov ie on-1 de
nd and dfu) Such bhot dlw is minimum anel uev-s

aeld u lo Cbr j o on-i)

s Slu]
tor even P e V-s olo cie eY V=o tw n-

divi-d Lu) +wluu)

end P
end hv

end
 3-3
Anays's
h-

V o
s eVe Vcev -S
ind t ad dlu)

n-) -04) (tn-.)- 0)

h-
2 an

anln-) 3

an-an A2

Tc, 9 (o)
V is dhe no
Veyce n»lv) S /

Sone ebe
3 decbnan votr F

Flay43

TYee veskces Remafnig Vevres I usakons

al-0) bla3) c vð d(a.7) f-, ) a-b 3 (a-

bla ) clb. 34) d(b,34a) el-,)

R-dS (a-b-d)
d (b.s)
clba) eldi5+a) Avo) -
a-c-y (A-b-c)

c (b4)
eld,9)eti - (ab-d-e

-e 9

eld.a)
find the shorte þath wm Source h desbi'nakorn
Pans desbnabon Ve«tex n he parame te alon with path matixp

Tnbalze mati P Cwih Soure vetex

Afte addtrg u o s, cke whahe u 1s deirabon

If u ts desknahon tenminale be r loop

inally p¥nt bhe path ling path matp ard he shov tes

dstare. 4S

S 20 3 S
IS o
30
a0 S 3S 'S
O

3 0 35

dlo) o
P)-5 5 5
d Li o 4
PL3 S 3 3 3
d23 a9 2 29 Pla S
d[3) Pl3)S 5 5
dA) = 39 34 3 4 3 P4 S 3
dls) o PLS) S S

S V- Ss dl)min ( dlv) ,dlud 4wlu)) PW-u nd

ftnd a,u, dlu)
Su ch dhat dl) is
munimum and ueV-S
0,),a3 ,A 3,4
5.3 0,1 3,4
PD-3
dl) 14 , 14
PL4 3
d[A) 39
513. 0,2,4 dlo o Pa)) 29
d) 29 PAJ=1
53,,a O A
dtA) 3
d L)-49
,34
d) 3h
5.31.2, d[o)=49 0,49
S.3,13,.
 b =4° 3-
Shovtes d'stanre aom 5 o

5 o 3 s decd oidh dstanme k. 3 to is See teel wdh

tance 4 1 0 14, I b 2 Sele bed with distance Jh 4I3 = 2 9 .
dstance
Vertex A is Selected in PLA, it is noE
Since there s ho change
Considened So no change in padh a 0 S S-leeteol wit ddstan ce

path 1s533515-5 6

Algodtm
for i = o bo n-iodo
dt') cost [ Sounce,)
PE) = Sounce

SC3 o
end fa
Sf8omce] = ]

pOY I o n- de
and dlu) such lhat dlu) s minimum and u ev-s
1ind u

ado u ws
F (u = - deabnaHon ) break
oT evem VEV-S odo
P dlul 4co f u) <dlv])
dlo) dluj wlu
PL- u
end f
end fov
end fp
 he hote podh rem.Aouee bo doskrahn
Hlgoatn h s t print
i dosirnahon
ohle 80mc
Conte i
pti)
nd l e
Onle
sou YCe,'= ' , d doss'nabtn

arcd dheu hre r

Diksra's algoibm Compones pah lontha

ads edge wsg
pi's algondhm Compones he eolge weAh agven.
 3
Op malre problem
Hufman trees and cocles

H/man algovihhm wa dev eloped by Davd Human

Hean
h algovidhm is basicaly a codia lechndq ue or
Encoding dota. Such an ncoctod data fs ucol in dala Co mprex on

techagu-
Algondhm.
Step Ini bale h onee hocla rees ard lobel them wib he
each characten in
Characle dhe alphabet. Rocored he kequeng 3
its re's ot o indtcale he trea's weght
StepRepeat e blouing openakon ank) a singta re is
obrained Find ho re wth be Smallet weishks ( Bes .can se
brolcen atbibrani'y) rh
Hake bem he lest ardåubtrea a new ree
Gnd recbrd be Bam
dhe we'ght in he o r of he new
ree a i weight -
ree Consuced ing dhi ag Callao ufman hreas Zt

Called Hman Coole

n dhe manner

Con Stle he Bue chara cte alphabe 0,8,c, D, - wi he

ol bwing Occunere Pnbabilikes

characte A D

O-1S
Proboblih 0.3S
 Oas /

Kesuakng cook wo eks

chavacte A D
Probabliy 03s
Co obe wovd
10d 00

DAD is
en coelesd s

T00l1olIolllo is dicoeo 13AD AD

with e Occunnence pmkab.like gven ard he Cacketoore
Xenats oblained, the expected nunben b pe characs in dh
Co S
 3
alo 35) +3 (o. ) +2 (o *3) +2 (o-9) +3
(0-is)
a as

Huffman's Coca ac hiom lhe Cormprebn rako. Cornpresion rako

sastanolard mene Compreson alnp vthrn pchene

ea t wl e 357. m
Huttmans enodin
an i tpsed deng encodirg
impgrtant le
Hurnan S encoir1 One he mott

Compresion methah.
In addkon o 4 Ainp)h and veYsebilihg, tpelek% an

pmase minima lergh encooir

Cockaiod omeeenee
huffmon S not m i o data compon
algovidhrn ana

Suppse bae n postkve rumbe u wn that

have to be asgneo b n keon binany re,onepe hedk.
Delne Ehe wezhbe pathlegth as the um, L:w;

-Deradt. of be Zimple padh pom he root o dhe e a

Cone truct a biray tre ilh rirrru

Ih poble Orna mary2ibiaion ínvolving deesion maling mali

E Jorrale,

The ame u esir achosen obje

Pocia b 0lrg u e b r s are wenahte
h o . D iferen
ctraleg pio Cha arne C be moolekd by deion
 H,
h0
(n:h)
(n-4)
no no
(n (n»3)
ho ho

( -
betweon land
dacision re v gusinyg an ege
Iwo
he vook bo a
le
of the Simple path om heelbd g e t
eength hekd
uebon
4uehona
inSuch a tYee seGal to dhe no 6
ted
dhe
l
to Lhe Chosen numbe. Yepvesen P,the 2am
8am
P the
chosen w'th probabi lihy
Ihumbe i 's
drm the rco
the S'mple path
2P: where : t s dhe lengh of
" 2ueabsons

to the h Jees. îrdicala he avera

averag
epYeen teol
ens bhe chosen hum be
wih a Game Stratene
by deciston ree
chavactes Prom &ome
Tolo encoole a bext that Compise
0f the tex t's
b ach
n-characte alphabet by as nng use
coclewoved,
bits Calleok the
CharacterS Some sesuence
bibBtrn
asnabo each characera
ysedength encona
Ofthe8ane engdh m
coda cwo rols of Eren
Vartable lengdh encodltng hich asaar
|lengths o different characters

Problem, huw t lell how i s of an encodao tex

many
Yepyent the S Charac tes.
 To avo h Complicabon, rnits b prek e n Cool.
31
n prek Cade, no Corleroovd saprefi of a co dew ord

anohey chavactev

encod A car a bt shring ArbiEo

yal 1s a C de worcl D or,e
he niup b i s
ha cha rac ton erd penl thi
chaxac tes place hese s by
is YCachecd
end
Openirn (nSl the bit sbming's

CYeate a binany prek coole tor 20rme alpha bet, ansocale

a brirany tre ir whith

the alphaberts Chanac ters with lea ve
and all dhe nshe elgp,
all the left eolges ane Jobekedl
av labeleo biy )
the o
recovorg he e
lahel
obtoineol by
Coclewovd
Cooleword aChanatten

the from the root o the characten's ea

Orn Afmple path

ranshym _and ngue apPach

WUT) as two sage proceoluns

instante
tanc
tbhe problem's ins
s In
In the trans brmakon Stage,
anothe, moe amenable
reason or
one
mods Red be. foY
to0lukon
and Conguengstag it ssplv
Three vaniakov

)Insta ne Sirpkcabon
Reprenenabov Charae
3 problemT eolucli on
 Ins tarre Bimpikahon transhvmakn-

bo Bimple 0r roore
Convenent insTane e 0f the Aame Pvblemn

Evans hvmaon to a d'ffe vent

Repvexentakon charge
vepresentaion of the same irtance
irgtance a dtlerant-
Problem reuckon tanshormahon to an

availab
PDblem
Poblem DY whic h an
algoithm s alrvaoy

Simpler inatanee
Probes lukon
inakance anothev representabon
DY

another ppblem'S ire ta ne

Heaps ard HeapSor

ovoloeol thot is suitable fo
Heap
Heap s a pavbally Ds

fmplemening pviovdly 9ueLO

AHeap Can be oefineol as a birany ree
e wth kaus
e cwith a
Heap
Ostgred o its noote (One koy per noele ) proviolaa the nocw ing
a conolifons aro met

The bree's Shape vequivement - the binany be is esenkall4

all :t levels full except possib he lart u e l

Cormplele ( ano

Whee Only Bome ghtmast leaves may be missi

T h a parentas dorrinare veguirement - the key at ead nacle is

Puakes than o eual ko Be kous at its chlben.

 lo 3-

Not a heap
Not a hea
Heap Cee Ahape requtrement is violabmd) Pavena dominance
eGurement fasls Br
Bhe hocle wrth koy s)
Koy values in a heap ave 0volenaol op down e a
Se_uenee values on
any pan om the roo o a loo 's deeveasira.
Propenke e heaps
.There extsu exactla One exsenally Complete Br wth

n nocla Its heht is egual bo L og

a. Root 7 a heap aluns contatns its lorgel elerrert

cons'odlorees cwtb all t oloocenodant s

3. A nae a heap
alo a heap
. heap Can be impkmentesl as an aay recovoling
t s elemmt in be kop olnon, lee to *ght a h l o n I t t s convenient

to sore te henps elernent in poston dhsough h uu

S ooh
h »
an

anay keavingH[o) efthen uruwesl oY puting dheru a Senbinal cohose

Value 1s 9wake Bhan every element in dhu heap

Hecp and i annay epresentakon

1o

inalex o 34S
6

Vada Pavents eaves

 a) a pa vental noode loys w11 hbe r t L% J
bhe amau. wkle &e
len, koy sl orrupy dhe lont postkor3
1 pos'hor
b)the ch:ldhen a koy any porentos poshor
Is "% ) wn be f
pos:konn ord + | . Tho
pa vent e
koy
in
poson ( ; <m)w1) be
pos: kon L/s J
Dehne a heap as on
elamant in posthon Vny (1... n) n isch evey
be sthol, bhe anny
in

hon
uoke
eGual o be elarment in the postkon and 1t1
H 2max
Ha:3. nl7+3j r i: l -
..

L'%J

AlgoRtmheapbmue ( H
Ci.nl)
fo L°l J don 1do

VH[
haogple
whe ro heap ared 24 Sn do

iF jin hene ane ch:ldsen

heop true
elne
H[)Hlj)
HLI:VV
 gotkoo up Conabruckon a heap brv be st a,9.9,6S,8 3-

()

wtb thi bottorn up algovidhn, a heap 8ze n can be

Constructrd wfth euwe han an CompariSons
Max hiap Is a bee in whsth valus , each noele is
qeaken than
0Y
eGuas bo the Value i children hocles
1

Min
heap Is a tre in whrch valuu 2 each hocl 's Los dhan ov

euas lo value of ts children rooles.

(20
 The altenate ard les elkeient
ala conaucta heap by
alg :

Succexstve inekom a
ane Key in a
preurouly onsnuctrd hanp
call it bp dam henp Conmuchon

insent l0

Delete voot ke kom a

heap

The inekon opeakor cannot reguire more k

Cormpav'sor han the heas hetght, since bhe hesht a heap

rth n
hooles is about losp ,the bime efkcieney of ineHon is
in D legr
 Heapsovt
Dis coveneo by Ja wtlam
Th fs a
hoo stag algovdh
Heap Conbuektn; cOnanuet a heap kiv a a n

Maximum delekonai Appl he o t delokon, 0perokon nI no to de

Yemaining heap
SovHng dhe d 8S, iS,3s,95, Ju5, 55, 165, 7s by hapso
Creakon )max hoap

as (S)

(14S)
JAS

(165 165 165

delekon o voo

(14S
(3

(16

(as

-/@

'S
 3-
ine Comple s:
heapiov n OlnJogn )
Iplaca orhrg algorth
Nor a Atalbe sovina medhool

7Enso Yun more Alow a n 4uthsont beraue Torcen p

Du is compekve wih meg sove

Masi mum y delebon om aheap.

with &e Pasb kay le of be heap
Echang
Echang bhe rOot's key

33- Heap
Heoh themalles Ere ahifkrs k odoun the ke

kame ng botem p heea

acbly in h

Conatruckon aia
he etciency delskonts determines by the no:) ka
Co mpantsons hasdad o heap:A he tree te bhe u a p ha

been maole aro &e brea s decrwoed yl siee

3p e&e
itCanrot Yequfre move a Conpanisons th a t twice he

henp's hetaht, he bme efkctency delekon in

o(esn)