ON 11-1L

Sytlabus
Tntnaduchon Aigonithm, pofonmance, Analyiu Spau Comdkexity
Asymphohc Notahion 09 oh Notation,,omega NoBation, Thete
Notakon and JiA oh Notakion
Divide &Congune Genenal method, applicahon - Binay Sanch,

ouicK Sond,Mog Sobn, Sihassen's matnt mulhplicahon

Alggonith:
An algaoithm isa inik setd insnuchons hak Can be
uSed to povfanm a pokculan Has.
An algo nithm should ponougs hollouwing chaactoni3hc3
1 Input-
h e in put d 2no an mane numben db 2uantihes Should
be 8iven to the
alganithm
| 2 OutPut
|AS an ouBput, alganth should Panoduce at least onu quant
3. Definideness,
Each insinuchion in an alganthn should be Speeidic
and unambiguous
4 Finideness
Thu algonihm shouid be sinhe. That means atoy inik
numben db sieps Should be dominati
S Etfechveness
Evo skp do algoithm shauld be easlble. In athu
uxohd3, euoy SHp db algonithm should be such hat
it Can be cnnied out pen and pencil
 PPho cess j00 Design and Analays iy ds algonithmn3

undesAnand -thu pnoblem

SoluHon as an algonitho

Algsnthm Desan Techniau

pnove CooneelnessS NO

Analyse thu AlGoithm

is it eK cent NO

Code he alganlthm

TPesd Algonthm
hou au pun yP dk alganthms
APnOximali alganithm
2. Pnoba bili8hc alganithm
3-3nvinik algonihm
tHuiShc algonithm
PPn4gimai Algonithm.
is Said 1d appnoxi Mat ib i 19 mfini
An alg0nHhm
Ea 2 -1. , 7-3.14..
epeakng
Pnobo bih8hc Alganlthm
uneonain thun i+
t h Solution dk a problm is
Called a ponoba biltShie alganth E: Tooskng d% a Coln.
 inginik Algønithm
is called as infník
An Algosnthm which is vb? fmih
algonn E dusion by 2ene, Complek s8) do a chessboad
Heunisie Algonithm
IS (allt
ving dewon nputs geHing mase oulpuls
the Heunistic algnthm3. ea All Gus)ness ApPphcakons.

TsSues bon Algoithm

I How o devise alganthm3 9
2. How to epness algonithms?

3. How to Ualidat algonihm.32

4. How td analy
alganBhms ?
S HOW t test a pnognanm

Ponfonmange Analysis
Porbonmant Analys nebos do. thu Hasxd delonmining
th ebicrengy an alsomthm ve. how much Compuhog time|
and Stbage an algositho) ezu i s o Mun 0n edecuk.
This analysrs algoalgosithm helps in judging rihe value
an algonthm ov anothun,
To guda an algoaivha,Pmiculab tuo Hhings au daren
inh ConSidoahon.
SPace compleni ty
2. Tme Compleny

spau complenty
The space complery an algsathm Cpnognam) is-ihe
amound Mumany it eeds to nun tD Comple tion
 The sPace needed
an algenithm has hu bollowing
Component
.Insinuch on spau
2. Dada Spac
3. Environ mana Shcck Space
Insinuchon spac 1s thu SPaCt needad o Stnod hu compikd
Vision dk the pobgnam mshouchons.
Dada SPaa is he Spaco nesded to Shoe all ConsBand
and vaniable ualues
Envinonment Stack Space is used dunin8 execuon 0
funchong
Spaa Complexity SC)= C+Sp (înstana chanedonat

Total ShPs
Algoitm Sum(a,n)-
S.0
n+1
tO i-1 n do

netunn s

Totad 2n+3

h u above algonthn Spac complua is 2M43 Sps

 2 Time complnitT
13thu amoung
ihu ime Compexi d& an alganithm
Gtme Complle time H needs t un t Completi an.
we Can measune hme complexity s an algsnHhm In
Hw 0PPnoaches
.Ponioii analas 03 Compile hme
2 oshenioni analyais 01 un (Exeeu Hond ime.
I n paioni analsis befone hi algonithmis eecukd we
Cwi analy« the behaviou d he algonthm. A prioni
delemining he onda do
analyris Concentnak3 n

ienecukon d& staBemends.

19 eMecUkd
I n Positn10n analyis while the alganhm
we measue he execution time. posten1oi analsis
J1v acra Nali values bed4 1S veCostlj.

Time Compenity TCP) - C+TP

Complexity d hpnithms
1Besl CaS
Inpuns au Pnovided in Sucha Guayhat the minimum
time is Aaequined t pnoCSsHhem
2. Aveage Cage

The amount time hu algoh akes on an avenag

Sel ds inpu.
3. CwOnegt Cast

Thu amGunt & time the algoithm taru on hu wDst

POssible Set ds Inpu.
 ASHmphhc Nolahon
Passible 1S
7ACCuati meaSuumend d time Complenity
Cwth asym phhc notahon.
ASmPhhc Nolahon o Complenit ivc an idea db hau
3apidy he SPac eguviemum 0 hme nenuiumunt
C
3ow as pobem SIe ncneses.
-In genOal asymphhic Compleaity fsa measune d
alganhn no pnoblun
an algonithm is
an a
0Suallythu time com plexity dk
numbn d
Ounchon elakng hu inpt ngh to thu
Sps 0 3tbnag Jocahon.

Theu au 44 mpoHHam asSumphohc notaons au

I.Bg oh noBahon (o)

2. Omega notaion ( )
3- Theta notaion ()
4. LiHl oh hotakon Co)

1B1g oh noiahon(
to
denokd by O. 4 15 ud
Big oh nodahon
is
to
descubedhe eicren d algosnithm, 34 13 uSed
epne Sent thu 8uPP) baund d& an algsishms umin
tim. USng B 0 noBahon, ue an gnve Maniun

amound db Hime akun bj tht algonidhm to Complee.

Dhinakon
be-the dws non-negotnve Aunchons
Jet -fn) and gtn)
s Said to be OCgtD) 1t and onlyiR
CDe Say hat fen
 hene exists g Poslhve Consanc and h Such
hat
Cno C*geno on al non-negahve values 0 D
when n 2ho, gtn) is Ahu uppn bound ndny

|fcm-OCgn

Thu above
deinakon Slaes hat the funchian f is alma
chme9 the tunchon wh1n ' is neakon Ahan dn
equad do o

p.Big omega noBahion

B19 omega notaia 1 denokd by . I 15 ud
epnesent the douw bound d an algonthn Junming
ime. Using obig Omega notadfon we Can gve Shanht7(oininu
amount d tme Hak1n by he algoythm Complee.

no
 Oc tinakon
h e tumction ferd Asd gn) be the tws non negoa hve
funchons, we SaH that dn is Sad h 6cgen) i
and onlyît thou exisis
a Pasihve Consam c and n
Such that,
HCn) z cgco| an all non-negahve valuy dk ®, C
whu nzho , an) is dhu Jouu bamel to

3. Big Thela Nohakun )

hu bg heta notahon is derokd by 8. 3 is betucen
uppouppon bound and Jowon brund d an algonithms
Junning hime.
Denahon
Lel fén) and gn) be thu tws nan-negahve funckons. ut
Say hat -ftn) is Sa1d br egn) it and only It hvu
enishs a pasihve ConGanis and Such that
Cgen 0 Ca13n on all non -negahve values
n,whuu nzhoo

ftn) OlIn)
 4. L144k oh Nolahi0n Co)
| JHHe oh Nolahon dnokd b 0,he osjmphvic upn
bound pouided by 0-roHahion may an maj no be
asymptuhaly hg. th bound 2n0c) is asymp-
Mcall oight, but the bound 2n0Cn)13 n
ue use 0-nolah on 4ho dinoe an upp bound tha4 is
no4 asymhially igh
deknakon
A ) -o(gen). 133 4m 4c+Jn|-dozn anj Posihve
ConGtants C>0, ng 0 and n>no.

Divid and Conguns method

Divids and Con2ue u Rnsbabldhu besk knswn gnUiad
olganithm design deehni qut. hu painciple behind the Divida
t s easien
and Congua6 alganithm design dechnigu is that
to Save Seunals malleh instana db a penhulas) ponoblum
hen thi Jange ont.
lhe divds and Congusa tehnigue nvolves Solving a pohtulab
problem by diuidingt int one an mone Sub pnablums
do Smalln sie, ecunsively Soluing ach Sub ponublem and
4hen moging thu Soluhon dk Sub pnublems to rodute
Solution do tha engina pnobem
DvIdi and conju) cons 3 phahs
Divide
2. ConuHe .
3. Combîne,
 O1uid
Divide hu Pnoblem in a numb ds Sma1t Sub-
Phoblems idally ds abut the Same Se.
2 ConguMAN
The Small Sub- pnablems u Slvd, +ypically
neeunsv
the ub- pnobkm Sr2w amu Smal enough, jus3,Soluc
the Sub- powblems ina shnighh Bonwarud manno
3. Combine
I & De cessony the Soluhon abBained tht Jma lln psublens
ane Connecded do gut the Saluhon to thu ignal
Ponoblem
|hi Shous DvIde and Congueh
tollouwing diagnam
tpical Caks
Paoblem
dvds

Sub pnobltm Sub pobem

Solve
Solve
Con2 u Sub fnoblem Subpnob/2m|

Soluhon t Subpno blem Soluhsn to Subproblee

camblne

Soluhyn to
Pono blem.
 APPliahons d& Divide mdCOngu) +

Binany Seanch
2. Quick Sont
3. Mhg Spnt

Binaas Seanch y

Seanch is an ebkicient Seanching deehnique that

Binany
twOnKS Hh only Sosndto J . So thu Jot must be Stmkd
done using, thu binathy Seaneh method. Gnaty Sanch
iS based on dwid and tonguu techniqu.
Ih pnoun Binang seasch is as tollaus
Thu method Sants euth Jooking athe mida'le elenun4 d
he Jrg. 1 mabthu wh hu ky element,hun tawth
iS Complek.
OhowiK ,thu uy element may be in the hihst. hal az

Stcond halt db th u
ha y elemund is Jeh han -thu middk elemund, hu
h Seanch tnhniaus wth the fiua halt d the dh9.
h e y elemunt freatn than th middle elnusa
thun th Stanch onhnióu uAh tha Suond halk t thu Jt) .
|This froa continw unhi) thu ky elemund rs dound af
th Saseh Jails mdrcaing hat hu Ky 1s ne theru.sn
he s
Eanpe

O lo112 13 1
Keq elumint A
,45,6 8111121 y
os|lo1812|3 93 1e|41 250 S0
Mid
m
Jow+hh) /2
o tu) l22

R
I0 1& 2 333 |e
halb
5
u
md elumin =3 SeCond halk
Jow So T32

Mow
|STo
mid(gtL) 12
J 1 a mid 3
ngh
so 23 3 8 1 1925s
hal Condhalb
mid elemuna 5 So 5-
Hhigh
6 6 00101 11 1
0 Sto 22 323 1 419s
m+d mid ow thih) l
Jow mio hrtb

S|1 18 23 3 9 o120
Hen the Sunh Kuy -1 i bound as paltion
Binay Seanch Mg09nrthm
Algonithm Gmany Search (0,n, k)

Jou-0
Hgh-n-1
whiu ( Jouw Lh1gh) do
midlowth:9h) |2
TsCKe =almid) hu0
heHusin md
elk ir Cud < almid)
High midti
el Cxey > alnid)
Jou mid+

RecunSivt BinanySanch
Algonithm &nsahch (a, n, kiy, Jow, high)

(Jow high) hun

gid (Jow +htgh)I2

18 CKy almd1) hun
etunn Mid
el 1 Cy almid
6tnseon ch Ca,n, Kr, Jolw, md-D
elk 1 CKy > alndD
Binseaneh Ca,n, kun, Jnda1, hgH)
Ltunn -1,
 Advantagy,
h maln
aduantag1 tbinj Search 1shad HS
bashen han Jnea Seanth
DIsadvanta4s
Thu man
duadvanBag obnay Seanch is hat Can
be applredo only a Sazded Ji1a d elemenh
Ginan1 Stameh IS unuedu unJupfal 1 th bn is unsorko,
Ehcrency
BesA Can OC
Avna Case O Clogn)
wonACa Ollog n)
Spatu Compkaity otn
QuicK SONt

h u 2uwck Sed is tonsicdned h be a os? method v San

he eemends wa devlopd by CAR Hoana.
This method 1s bastd on duids and Congutn dechniiuw,
ie. the enKsu J * i dvided i into vanvous Panhihons and
Sonting t applitd again and agam on hox panhins.
Ihu method aho called panhtin achang Sast
examl 62 8o 82 60 s1 s1 42

pvOt
 PIVU

42 62 8 82 601s

phuo

4 5i 626os&|
Pvo

pluon

6o 62 880 1
Subli 1

inal
42 5 2 60 62 8 0 2
Alonithm 2uicrso (Jotw, high)
(dow <high) the0
K Patthion (a, Jow, high*)
uICKSant Jow, K-0
2uickso (ra hig)

Algonithm Panh hon (a, Jow, hig

pivot aflowj
Jb Jow
ubhigh
while (Jb 4a ub) do

whie (a [lbj<Pvoi)
Jb Jb+1
whiu (a CubJ 2 Puot
ub ub-1
(Ubiub) then
SwaP Ca dow, high)
alJou3.a a Lub3
a lub z pluot
netuno ub

Algoonthm Swap Ca,3)

Htmpcati3
atii aC
a Co3mp
 Bes CaSe OCnlagn
Avnage Case

wons ase olnt)

SPaca Compleaitg OCnD

Randomited Quick Sont Algonithm

uick Sort is good an avOnage, but bad on centain uonesi
Case instances.
FRom above mentioned hhee choicos andomi2ahon 13 the
beH Solukon to impanave dhu abonmana d quick Sod
This is because by having ondom plvot Selechon we ane
imposi ng he equal
pnobability d govod Pofonmana.
Henco andomizahion is a tovl to mpnove algoni-¥hmps
wHh lood wons-case but
enenal
90d avonage Caße complexity.
hu oonSt- CaSe is Shll
-hoe, but lwe almat Centainly wont
Sea it

Algonithm uicksa.R( Jow, high)

1 (Jowz high) then
i& (douo <high) » s) hen
Swap Ca, (naodom(mad Chigh- Jow+ D4 Aaw), Jou)
: pohhon (a, Jow, higha0
2uica Sovd-R(Jou, k*);
2uicK SondR CKAl, h19h):
Ma SonB
S o n d 1S based on divide- and-Conqu dechn13ue.
Men Sont method tonsist two phaaty Dviding i Høging
Phase
l. Dividing phase

Duing he dividing phase, each time the given J} d&

elemends is diuided nto two paHs. his division ponoa
ConAnues unl the dist is Small enough to divide
2.Muging Phase
l o g i n g is the pnaass ds Combining two Sankd di8193
So that,the esultat Jisa i3 also the Soted one. A iS
a Somed Jis wth n elemet anol. B is Q Somled di} with

n elemenis. The openaion that Com bines he elemen1s db

A and ind a singe Sonted Jut C wtth nntn,
elementes S called maging
Algonthm nege-sa( doto, hrgh)

1 Cn 1) thd
netunn
else

i(Jow high) then

mid ouw thsgh)I2
mee- Sd (douw, mid);
mag-3nB Cnidal, hgh)
Combine CJow, mid, h1gh)
 0
Ea det the Aist iS - 500, 345, 13,256, 98, 1, 12,3, 34,45,48, 92

Soo 345 13 256 81 12 3445 a 2

suo 345 131568 n 3 34 45 |8

Soo 34 13 45 8 9

Stro
34s 5 1 3 9

13 345 Sr 98 25%

1313 98 2563495 3 12 34 45 28 2

3 12
13 34 45 a8 92 8 256 34s s
Sonded Jist
Alganithm Combine Cdow, mid, high)

Pdow

Jmid+1
ohile (upl: mta) AND (j<2high) do
I CatuP <z aLii) hn

HemPt: alu
up up+1

else

tmPC13 aLiJ

(up> mid) theo

fot to h:gh do
emp Lij: a I ;

else
fut up to mid do

temp Ci3e atI;

fon tdous to hgh do

atty mpLtJ,
 uonsi CaSe
Bes CaSe
o(nlbg
OCnlogn
Avonage CaSe
|Ofnlogn)
Sace complenay OCzn

Hassen's laiox Mulhplicadion,

Suppase we cwantt mulkply two malniee Aand B ach
ob SI2 Al i-e.
C Ax8 thao

Ci C CAn An Bn Bn
A2 A 821 B22
Th muthpliabon 9iv

C Au xB11 + A2 Bz
CILAu 812 + AnyB2
C21A21xB1+A22B21
C22A21rBn + Ai2 x B2)

Thuso accomplish 2x2 madnix mulhpiation these ne

tolal mukipliahms and 4 addihans
8

Algonithm Mat.hal CA, B, C, n)

1
fun i:l ndo
An a o n do

Time tomplexity 48 above algontm 1s - O(n>n+):O()

 Can be
Snassn showedthat 2x2 maBnlx mulhpicahion
acomphshed In 7mulhpication and 18 addikons 0)

Subinacho ns
hhu Divide and Congue appmoach Can be used ton imple
mening Sinasen's madnix Mulhplkation.
IDivide
Dvide. ma1iice9 Inb Sub-malNiCeJ Ao, Ar, H2 - CH
2. Conguen

Use a gnoup db malshix mulhply equahons.

3- Combine

RecunSively and get

get the
mulkply Sub-monites he
inal aesult d mulkpli takons aHn fofon ming
Teuined additions 0 Subiachons.
C Cn An A2 B B
Au An2
Ca C22 Ba

S A u tR2a) (B 1 B1
S2 CA21+ A22) x Bin
33 An X CBi2 - B22)

S 2 A1X (B2) - 6n)

S5 An 4An) XCra
S6 (Ag1-An) zBu481a)
$a (Aiz -

Az) X (B2u +622)

C S1+S4-55 +S
Ci2 S3+55
C21 S2 + Sy

|Can S t S3-S2-+ S6
Now weill Compae th actual Oun nadi ional matnix
mulh pli cahon twith assmen's Pnobdusu
C St S4 -S5+39
C n4Az) (Gu+Br)-A22 x(BuBn)-Aut Pi) x Ba'
+(A-A) xCa/t B32)
C Au Bnt An Br

Alganiahm S mul ( na *A, in +3, In ¢ , n} n)

1 (n-z1) hen

elst
SH-Mul (A, B, 4, n/4)
St-mut A, B+ (n/4), Ct (n14), ni4)
St-mut A2 *Cnl4 D, B, * (nia), n
Ct2
S-mul lA+2 * (nl4), B+hl4, t+3 *(nl4) , nlu)
St mulA+tn4), B+2(ni4), C, n)4)
S4- mul l A+
(nl4), BH3 (n]u), C+(n/4), Nl4)|
SAmul CA+3* Cni4), B+2* (nL4), Ct
2*Cnls), h/4)1
6 mul (A +3" (n4), 543* CnJ4), C+ 39(n14),
0140|

Time (omplmity Tn) 2