0 ratings0% found this document useful (0 votes) 155 views34 pagesDAA (Module1) Text Book
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content,
claim it here.
Available Formats
Download as PDF or read online on Scribd
‘Algorithm ?
pod veo
ia
fe sequence and such an algorthe, PO
seri ft ant of ie Mm shag
igor: TH
“eto of
Conecrest
a [| on
exttany
Fig 114 Natlon of lor
KEEY Properties of Algorithm
Sergiy wring he sue of insrctions as an algorithm is not suet to
conpsh eta ts 1s recesay to have folowing properties assocted with w,
end
1 Nowambigty : Eich step in an algorithm should be non-ambiguous. Tt
seas ech insrucion should be dear and precie. The isiniction in an
algonm should not denote any conflicting meaning, This property alo
india he efi of ago,
1 Range of input : The range of input should be specified. This is because
rox loth is input deiven and ifthe range ofthe input is not been
‘pected ten goin can go in anit state
4 Maly: The same algorithm canbe represented in several ferent ways.
‘Tht mens we cn wit in simple English the sequence of instruction or we
ante it in the form of pseudocode. Silay for solving the same problem
we can wnt several diferent algorithms. For instance: for searching a number
7 the given list we can use sequential search or a binary search methed.
fe “sacig” is tek and use of either a “sepuential serch metitd”
‘inary serch meta” san algorithm,
TEOHNCAL PUBLICATIONS? Anup hut for owledge
ang an Ast Aas ba
4 Speed : The algontns are writen using some speci ideas (hich is
populatly known as logic of algorithm). But such algoits should be efficent
and should produce the output with fs speed.
'5,Fnileness The algorithm should be finite That means afer performing,
required operations i sould terminate
1 arb
2, ti, Wt he ht oat mst ey? |
13 Wat om leat ? Wht re the properties fn algo” Exh on exe
Algorithm Specification
Algorithm is basically a sequence of instructions writen in simple English language
‘The algorithm is broadly divided into two sections ~
gor eadleg
cor fran cf ag,
‘eseon, mpt 8 op
‘Agonti bay
crs of gel aay he gern
bj mating ee aout ryan
ens and ssn Senet
Fig. 12.1 Structure of algothm
Let us understand some rules for writing the algorithm,
1. Algorithm is a procedure consisting of heading and body. The heading cons
of Keyword Algorithm and name of the algorithm and parameter list
syntax is
Agr nan p52 oe
Tis toyi@ sous Hore wre
ewntentrs enamel oaamters
erage any)
TECHICAL PUBLIGATIONS® An opt wedgeLab
nets AOU
—_ ve st write followi s
nae many conve tt we loi i
"rten DS?
Imes
(rom:
3. The tay of a a
Betti whe or one
reaped tenet sl bee
Sedu comners oe ten ig big of men,
ae eer shold begin by Teter and ot by dig An Keifer can
ten ef banner Sing 2
verre wo write aaa Opes explicly for Hemost
1 eT at el a Spd gr, a gt
rejean and soon. The pointer type i also used to point memory locaton, Ty
compound data pe such as srctre or ord can abso be used i
stony aime oper © a ane Sere a Be
thn swe, in which vanous programing cong
asignmen statements may be written ts
sed within {and ] brackets
Res
For instance
Vavable —exprosion
18 There are other types of operators such as Boolean operators such 25 true op
fake. Logical operators suchas and, or, not. And relational operators such as
9, The anay indices are stored within square brackets |! T. The index of aray
usually start at zero, The multidimensional arrays can also be used in algorithm,
10. The puting and outputing can be done using red and write.
Foreample
‘wrt Tas moseage wil be displayed on console)
ral)
11, The conditional statements such as sfthen or iftherelse are write in
following form
sf condtion then statement
4 (condition) thon satment eee statement
the then statement is of compound type then ( and } should be used for
eclesing block.
12 while statement can be writen as
‘while feonitier) do
{
statment 1
satement 2
TEoACALPUBLOATNS® pt fr ope
While the condition is tue the block enclosed vith |] gels executed otherwise
statement afer wil be execute.
13, The general form for writing for loop is
for variable = vale, to value, do
‘
snatament 1
‘natorent 2
statement 2
)
Here value, is initialization condition and valu, is 2 terminating condition.
Sometime a keyword step i wsed to denote increament or decreament the value of
‘arable for ample
fori tonctep ¢ —| Here variable jis
incremented by 1
at each iteration
ite 0)
}
14 The repeat - uni statement canbe writen as
tatemeet 1
statemect 2
snaterest 9
‘unt (condition)
15. The break statement is used to elt from inner loop. The return statement is
sed to return conzl fom one point to another. Genealy used while exiting
from function.
[Note that statements in an algorithm executes in sequential order i in the same
order as they appearone alter the ober.
Some Example
‘Brample 1: Write an algorithm to count the sum of n numbers
‘Agotthm sum (2)
‘Pre, Description : Th alg is for ining the
inn of geen numbers
(Mop: murbers
(yOutpat The sum of m aumbers
reaul 0
forielwndoieiel
TECACAL PUBLICATIONS - Anup! rienaugstes — tebe
cae
rani 700841
ean agritan tock wht gen munber re ro
vats (rl)
eta rt ot test whether given
te a
ae ati
(a gem
1 asz=0) he
(raed eg ves be re")
ise
wrt ‘Given umber sof")
ample 3: Wit a lgorin or song the elements
Agr a (8)
an Deon: sortg th element is acending order
at Azanay ain wach He elaments af ted and 2
stot umber of lets te aay
now: Te sre ay
fericttondo
forieistton-1 do
(
am
temp < al
jab
age temp
)
t
Itai sna
Trap Wn atin i irl fn amber
tts)
os epi Th pin
‘Het given number n = —
faethe eta nob
fesse
foe ter cress be
tte en
a
-
“amen
sample agi fo plo mip of wo ma
ae MullA.B.n) : =
ae
/sultipication of two matrices Dee
‘isd eth
TEOINCAL PUBLEATIONS™ Aw
se
owtp Toe mulipbeatin esl wl bein mati ©
forie-ttondo
for je ttondo
hilo
fork ton do
Cty) Ci} ABKIBE
1 i pti 7h he eri of on hn
| 1 Wht isa gel ? Expl th nto of gent si a x
ample
4. fie an loth. Discs th crea of a agra tha
‘Analysis Framework
/ kaos
F
Sones
Fig. 131 Aaa Famowort
iency ofan algoriten can be decided by measuring the performance of 2
The effi
mance of an algorithm By computing two factors.
lgorith, We can measure the perf
1, Amount of ime required by an algorithm to execute
12 Amount of storage required by an algorithm
This popularly Krown atime complesity and space complet ofan algo
‘Space Complexity
“The space complet canbe defined as amount leroy eure ya
tenn
gorithm
TeemucAL PUBLICATIONS Anu ot noees!
—
we 9
re cmp pe cnet wt 0 FS = CM ag
train. The ace regarenent Sp) an be Brena “~
[secre |
etn Gal pan dates he pts ang
cr evo as
Ieee oro eve arcs TS ete pr hy
depen on ria pole este. an
rr ne vo yet of components bat cent © te space compet
pat an vale pt ~
Te led pt ince spe
Fem
+ Vai
aay sae
1 Spee for estas
Theale pa des space or
1 peat wae Se s dependent pen he parca pcblem iste be
sed The conte steers ich fe, do, Whe che) ae sed ie
such nse ;
Reus sack fr anding reuse
Cie oe ample of alpen ocemule he space comply
ample 1: Cenpute he spc gue! by olwigalgortn
ier A.)
(rr ncn agit cept aon
‘hate dont
(teat we cf tog
ops Th atin seed
rir vote
The spc reer oral ven in ample 1s
Sp)=c 8) =0
we ase th, bande cep one word se then sie comes ob 3
Eumple 2 + Conpue the sce needed by te fling agri justify you
bee C88
set Seta
sono.
Fast
cenpanddosrmaiagenes bt
tn the given code we eguie space for
ange oO
fous ten = On)
wines ath on)
fen” 00)
eee es camplesiy ot even ago can be denoted in ter of bigsh
roan 15 On
nample 3 : Compute the space requied by flowing agri,
‘mete Despina rocuseve grim wi
pete atic a he samen io a2 2707
‘Pag at of xt pe, otal be of emocts
‘ares ay
i sna aon of less of a8)
(evare Add tix
"The space requeren
0)210 + 1)
“the inurl sac sed for recursion includes space for foral parameters laa!
a a aoa dress The space required by each calf function Aad requires
ae ends pure frm va pare fr rer adress + poner to). The
Da SSaon ef nes all 1 fon an one eu cl) Te ecusion
SER guce wl be = 30
Time Complexity
“Te tne complet ofa algorithm she amount of computer ine required by an
lgoithn to runt completion.
Tis ificll to compute the ne complexity in tenn of physically clocked time
A ey mula stem, executing Sime depends on many faces such 35 ~
For
‘System load
4+ Number of oer programs runing
+ Tnstracton se used
Speed of undedying hardvare
She ime complet i therefore giver in terms of frequency cout.
Frequeny count sa count denoting numer of times of execution of sateen.
eb Te sn cst et ec mayne Ps sae
For Example : Consider fllowing code for courting te frequency count
vel fant)
{
ta
0: 4
TECHNICAL PUBLCATIONS® «A ptr or ienpreesmnn tits
: ‘of above program is 2
ego!
De emo on
sein
ti
4
o =
ot
Posey a
‘ a
L =
pen
, The frequency ‘count of above code is an +3.
en ingen of code ened ime he te ni
hen the condition becomes false. Hence forthe for loo ,
suerte he frp wil Be eed oy he
this semen wil Be eves fy
The fe opin above
iste and ne more ie
fresuency count +1 THE
the conan inside the for loop i tre. Therefore
ae at pint Steet wil be exec fron
EERE ee et he iting
Selon
sei
tea
Bits at
ariastt) -nle+t)
ii: an
,
The feguency count = +1) +m 1) = +n +1 = 2 +m)+2
(EERIE ose ey ut fr te fing oe
‘rpeticeni+ +)
{
tetricene
no
ferferttc=nit +)
cil=< a be
eee
_ TEC Plano an op
wim tm in *
=
ee a
Se
tte =a
illite nin) al
$= —
[nine oe |
To a eaato ano
‘ater counting the frequency count, the constant terms cn be neglected and oly the
rier of magutade is considered, The tine compesiy is denoted in tems of
igotmic nouns. Te Big oh rotation 6 & most conmenly usd alginic
aarvon For the above Fequeney count all the conslant terms are neglected end ony
{pe onder of magntode ofthe plynomial i coniered, Hence the tine comply for
the above code en be Of). The higher oer isthe polynomials aways considered
Measuring a opt Sze
iis cieed ht fe pa ae ages en aly pet ra fr a ange
ee meat the een ac apt fant wich apt
a eed puter Seis oiler an egihn We eae pt
sae silg a inp ie Tor eal whe peomaing main fo mes
soso over of Wiese races Then ely we can eet the lene of
aoe wines te np a's len tan appoint Yale. Fx empl, a
‘Pal hing igre we con pedi nc he pt
Measuring Running Time
We have already discussed thatthe time complexity is measured in terms of uit
called frequency count
“The ime whichis measured for analyzing an algorithm is generally running time
From an algorithm
4 We fst identify the important operation (core logic) of a”
operation is called the base operation.
1 is not aif to identify basic operation from an algorithm, Generally the
aperaon which s more tne conning isa bse operon inthe algrth
a
TECHMCAL PUBLICATIONS? Anup it owt
gorithm. This=
conan ES
1.
eos
No such basic operation ec rample in
ss ne 2
pte cats i ba pean ‘The concept of basic operations
ea i ao =
Basie operation
CConpatison of hey wi
Lato dene
eer element ofa
Seung» tele
Pertti
omni oe Ae ign oy
eae tenets in the mace,
photon
copii OD ee
ses
Two mane
ble 43.1 Basie operations frm input
Then we compute foal
amber of tine taken BY
ts base operation. We
can compute De
rewing tine of basic muir! Sete
perton by given rel
Sonmula
«ing th fra te cpting tne canbe aed apposinatey
est Case, Worst Case and Average Case Analysis
an algorte tkes minimum amount of tine to rin to completion for a. spec
set of inp then tis call best case tie complexity.
For example: While searching a particular element by using sequential search we ge!
the desired element at fist place itl thn its called best case time complexity.
an algortm takes maximum amount of time to run to completion fora specifics
of npat hen it 8 cled worst case ie complexity.
For example : While searching an element by using linear searching, method if
esd clement is placed atthe end ofthe list then we get worst time complexity
‘The time camplenity that we get for certain set of inputs is as a average same, Then
{or corespnding input such time comple i called average case time comple
Consider the following algorithm
Agr So ec 2 Ltn
Problem Dessption: This elortim ifr searching the
‘Moy eoent tos an ara X11] squeal,
‘Mpa: nary X01] and search kay
TEOWEA FURIOATON. Ap ae owe
pt ase tm ne ey a pret
foicamsiae
tacbriiow
Best case time complet 6 2 time complty when an algoritan ras for short
tare, In above searching algorithm the elemen hey is searched from the lst of
Cdemens. I the ey element i present at fist location in the tist(0.-1) then
Sigorthm ran for avery short tine and thereby we will ge the Det case ime
amples, We can dnote the best case time complet 2
Chet 1
Worst case time complexity
‘Worst case time complet & 2 te compleity when algorithm ras for «longest
ime, In above searching algsibn the element key searched fom the Ist of w
aare ets the Ke elements present at location ten cel the algorti wil ran
fe Tonge ene and tery we wl gt he worst case ine compen, We in dente
the worst case tne compen 8
Coa =
‘Toe agoitun guaranies tht for any instanceof input which i of sizer the
roving ine wil not exceed Cua Hence the worst case tie comply sve
Important norman abou the ficiency of Sion
‘Average case time compleity
‘Ths typeof complet tes oration aboot the behaviout of an algritun on
spect on rndom pol Le ws sndestnd some teins fat are cee fr
“Computing average cas ime complet
‘Let the algritu is for sequential serch an
be a probability of geting succes seach,
‘isthe foal mamber of elements inthe st
Te fist match of te element wil ocur at {* loro. Hence probity of
curing ist atch is Pn fr every een
‘Th probably of geting unsucesul search (L-)
Nw, we can find average ase tie comply Cy) 2°
Col) = Petit of ice seth or deri 302 in be ist)
«+ Probability of unsucesfl search
ECHIGAL PUBLATINS® An op ft for netelements at whieh
chances of not geting
“Thus we can obtain the general formula for computing average cae time complexy
Suppose if P= 0 that means there is no successful search ie. we have scanned i
centre list ofn elements and still we do not found the desired element in the List ten
sucha situation,
Corgi)
Cong)
‘Thus the average case running time complexity becomes equal 1
Suppose iP = 1 ie. we get a sucessful search then
On+ N/2+n 0)
Cogla) + 1 (04 /2400-1)
Coglt) = (41/2
That mean he algorithm sans about half ofthe elements rom he ist
For caelting average case time complenty we have to consider probably of
geting sucess of the operation And any operation in the algorithm i heavy
Cependent on input elements. Thus computing average case tine complet is difical
than computing worst case and best case time completes.
KEG order crowth
Measuring the performance ofan lot nation wth he np sie mi cll
ender of growth For ample, the order of roth for vaying input sie of so
given below
Dean ahcnrascpating StS en
We wil plot the graph for these values
st
‘Fig. 132 Rate of growth of common comping tie function
From the above drawn graph itis clear tat the logarithmic function isthe slowest
"And the exponential function 2" is fastest and grows rapidly with
ales even for small input.
owing function
‘aying input sien. The exponential function ves hag
For instance: forthe value of mai6 we get 2° = 65536
Ts A
: ; ~ 2 1 De ied cont =
i : jl “ 4 2 Eaplnwith an eal Bw new sre coun inated 4 pga cnt se
i ; £ a find the numberof steps needed by & program to solve « particular problem instance:
i
a a a ee 5. vit ch den ranean etl ah Ds
2 ze ee 34567296 rt se, Bet case and average cs eceny of hs git
Tae oar ee
TEOICAL PUBLICATIONS? An petro
_
“TECHNICAL PUBLICATIONS® - Anup bust for oniAsymptotic Notations CT goa ee
1 ee) Cnn saad eps eine one
1 tang mympane roan we cn ge fine compost 2 “Taste pony
"ont pb or aera
anette such a2, © and © sed ae called asymptotic nations
HEED gon Notion
The ig oh ection dled by 1.118 a aod of repr Be op
sol etgaars neg tine, Dig bi ch sion we cn ve ng aes
‘of time taken by the algorithm to complete.
ein
1a and gb two cepa fons
| tn and roa cero lege su tat ng dence sone va of np
Lend 'n> ng Sle sone ea such ule > © Ne cn ete
ea
| ZA
| "
| :
——E
ecton)
14
Here fi) is big ob of gt is alo denoted as fe) € © (tn). In ater word i)
des than gfn if gis mull of some constant
Example: Consider function fn) = 2n + 2 and gh) =n. Then we have to find some
‘constant ¢, $0 that fin) $c gin). As fin) = 2n + 2 and gin) = n? then we find ¢ for n=
1 ihen
fie) = ane? =20)42
"EOWA FSU? A wat re
aoa Aes Ass ee
wet
le A) > an)
n=? he,
fn) = 22) +26
ao) = 2?
en 4
le fi) >
n= 3 hen,
fin)
ein)
ai)
ie) © gh) tue
ence we can concide that form > 2, we obiin
in) < af)
‘Thus alvays upper bound of existing tie is obtained by big oh notation
FEE megs Notation
(Omega notation is denoted by‘. This notation is used to represent the lower bound
of algthn’s running time Using omega notation we can denote shortest amount of
tame taken by algorithm.
‘Detrition
A futon fs sid te 2 (ff) 6 bounded teow by some posve
constant mul of gn) such that
fin) 2 c¥gi) __Feralln2 ng
Its denote as fn) (gin). Following graph ilusrates the curve for notation.
| :
e-)
ean)
Fig. 142 .
TECHNICAL PUBLICATIONS «An pa fr onenp mspnpsietbes _1' 18} Ite
sale:
amide fo) = 20° Sand ge) =7"
Then
‘)
tn)
20
=70
=o
fe fn)» aio)
nel
fie) = 207-597
st)
aut
=m)
Tie fn) = gio)
Une 3then,
fin) = 207 +5
= 4528
= 73)
2
fin) > gt)
Thus form > 3 we get fn) >€* a
1 can be represented as
antes 6 2)
Simay any
n} 6 Qe
KEK € Notation
‘The theta notation is denoted by @. By this method the running time is betwee
upper bound and lower bound
[penton
Lat fa) and g(r) be two nonnegative functions. There are two positive constans
namely yan c such hat
«sa(o) $4) Seto)
en)
osonntteesctdinre tothe
Then wean uy hat
fa) ¢ 4)
“Tea ratten K€ ®t)
ig. 143
Eanple
fy) = 2 Band 6) = Sn
where 022
Silay fo) = 2048
gn) = Jn
je Sue tne BeT Fern?
Hee = = 8 and Ge #7 wilhng
‘The theta notation is more precise with bth big oh and omega notation
Som examples of asymptotic order
1) logan) hen
Jog € Of) Jogan $ On the ode of growth fog is
slower than
logan € tn?) gan Oln?), the order of growth flog 2n
is slower than n? a wel
But
dogan € 210) logan $ 9) and if certain function Fn is
belonging to O(n) it shoul satisfy the
condition fin) > €* gin)
Tosca. PLBLCATIONS® «Anup rt nownh Apts 1-20 .
Dn anti pee Ea
SS cegnansénies el Apt
say gs € 20°) 010) ten
2) Let Mo) sin = 3/2 ‘Ths itech noon ty
bie 0 fn «et fran >
in-/2€019) fn)» Ot) we gt sehr mit ot depend on ba ped nc
s: maimarm oder? whch ID meme sate textes")
Hence fi) € Of) Ohh fein at
But nia-1/2e0(n") Astin) s O10) fin) = ann 4m a
snd nin = 1726010") 1 web 8 ot he th gun) becomes
Say, ae
mr /2e200) f200 toy = aSs o
nin -1/262103) fi) 2 i where Ais Beal
a= 1/240(0") fa)> 10°) 1 uation
to etn 9 = «(8 9]
‘As we know, O(n) i the of actons with eAletetecel
fa) 5 €> 80 = Ale’)
Then fis big ob fe). 7 oAote)
‘Definition +The Ts Tle oko & m8 appoaces Yon, We can wale hae —
| fin) = 0((0) when m9 eaaton 28
‘That mens fy) = AS nea 263 45)
fo) =
i) 9
oy gn) = Ao?)
‘This ako implies that) much smaller than gf) orn near ng 4-0(s*)
Fer example equation Cs
» 1 = o(s) whenn0 ty = abe? = A0(0)
Fy 1 # o(e?) whenn-+0 ca
3) Snt4 = oln2jas Tis fit) = An =A0(n)
fa) < c+ fis bvays te a
ae Thus neglecting the constant erm we will have
f) = O(a")
TECHINCAL PUBLICATIONS® «An up tus fr ronda TECHNICAL PUBLCATINS® A up trotad aya tA be _
=. oe
Pe nso, 920%) MRE
Sein Theta tn ees esprit Se
Sh Mt ie eh
Heme ceit
Fin) s cea)
Then Fn) € O1G in)
In short we wil ind he vats of ne such hat
Fin) 5 cog)
dssane Fn) = Sn? +20?
any en?
Then orn > = 2 and ¢=4 (Fn) Oh) swe Thats when n= 2 and c= 4
Fo) = Sate?
= sa)Pemnt
<2 LHS
soy = 0a
crate) = 408
=2 RHS
LHS = RISis ds proved
Bat when
Fe) « 3°
fio) = 2° then et us find
as co2"
te no mh ae of wh > (3)
}
word< (2°) Wl never be ru However 3°> 2" sivas)
Tt
Hence 3° + 0(2") In ob
TECINCAL PULEATIONS® A ptt onc
ang dee Ars a samsctor
ERIN» = OU) 007: (9-04) en swe
7 (0) = tm
Sali: Ll there be some constant c such ha
Tia) Seguin) forme ny a
Smarter wil be soe constant >
Such hat
Thi sags)
Le «y= mite) Soh ae as
forn2 ns a
a)
“Then we can write wing equation (0) and (2) as
y(n) Te) $c +e: Bale)
5 (e+e) e820)
5 56) i> 5210)
4 es Uma t)+ 83 In)
54740) Orang a) 53) 8 me
State ter he lling ere fie
rsnr3e (0), x? OUR, 17500442), w?steM0)
io
Solution: 6) n?+n45 € ln) uur Because for ig oh naan fi) < +
tr fg) =n a5 6 asume and gn) = 9 enn! on+S€0%n") i ue
Gi) nPs¥¢ ONO). Assume MO) = xP sT and gf) = 100. thn
fn) = c¥ ln) But +1 > 10M mfr any 921 Hence x? +1 £0.00) ae
(ays? +56 04) te because © naan represen verge ese complet
(iv) s16Qin) is true because omega notation is used to denote lower bound
ile. fn) 2 €* gin). Hence if we assume f(n) = n? 1 and gla) = then the omega
nih ets sats by thse vo feos
(QEERDREDD in te obra for heating
Diggs Li in nln i) 6re208 ot
Solution ) log n +i = O (n). Finding bigoh means finding the upper bound in logn
‘and Jf upper bound is vn ie. n!/®, Hence in terms of big.ch notation it will be Otn)
1) nm og In hs epson mg epee the pp our vale ence
willbe O(a gm
TEoAucaL PUBLCATINS® An ptt onDadi edna
10) e2at an! wi be O(6') By neging conta be ups ys
lyon in, Hencein ters of big h notation tie comple ety
“ ae
SID oe ny ec! ste nates)
‘Figures. Express the fllowing function using asymptotic notations Mista,
D6P on w/2Ke-1)
ize the necessary steps o prove the same
Solution : Asymptotic notations - Refer section 14
CEI,
)6"2 +n? Consider
Hi) = 642" 40? and
Bin) = 2" Then
6e2" en? © OR") because
For big oh notation
Hin) se gin) is rue,
Consider n = 1 and ct be 7 then
62% en? < co?"
fe 6924412 < 742!
Bs M4 istue
Similatly fn = 2and ¢= 7 then
62% 4n? sco"
ie 682742? ¢ 702?
Be ome
Thus 2" +n? © O(2") is true fore 2 7
uin-n
3 a1) 692 To prvi we ee ier F< Te
er gues pe Sand net) ods 0b pred
oman
Consider proof of upper bound -
Lip) « Untetng ta? for ae
falo-1) = Satins dn? for nz 0
Now, consider prof of lower bound
lew 6
1
when n=2
day
H
east
sa oi 160
Pre the €0(2")
) The acon 1 be 2 (gn
mana
1 fi) i Bounded
Soltion sn! <2(2")
tetow sme positive constant
gin) = 0
So Ko) = s(n)
my onel
ny = 2
sie) =2
wo a) # ato)
noone
fa) = Beaies
wy-8
so in) < ain)
Tas fo) < 809
soa! en(2")
ENE er its ins min ae
Setuton:') Refer ample 19 «)Reeeample$()
TECHOUGAL PUBLICATIONS? «An up tus for knows
ecwcas PuBLcATIONS Ana tt oea ntti Apoins __1 5 _—_______n
REED bic Efcency Castes
Meet ry many ma Bec pata
ne eat mir B08
axe ot on Mega ft of Goth rely yt MS
a atnd © We cn have diferent eficney caesar ech dats poy rt fans
et ete us the cation of ile ore of growth oe
Nine af Onerat Delon = _
ee = ypu de poe Pi | 26 = 16
Cont ey 7
ae hs comptes sow
ac Tce pt npn Pein 2
eo guacho ea — Tegan) < 9%
SETA tc soca
Dieter rater probes (ERED co te fot om
Ines —_—)
t— es —__—__—______ ston + For compro we wil coir aus vl of
en 5 ani Soran Peron mag
Pee oan oer
aera nals Stig be Seg i
= ne ae etofaaen ment eraabhon® Yqqeen = fe2a-0
Guna Ti When he grt has two Seaming mais lent :
wen See eraap 3
a
os = ||
rr Sess maken
ine oon. |
Speman eating Geeing nines) we
ee oor
rrr ——s—s
TaN Wn tis Geng al eo
Wb mens .
Srp ccs ,
ec Ifn=!
Tae 141 Bs empte onc nen Fee oho
Compare orders of root of log (1) and z 7
Satin + Far compan, we Wil enie varios vais of
wae 7
Heil) eee |All such comparisons indicate that
Lat)
fy =1
‘Then by sing ten elton we get)
forn>0
1. When n= 12.3 and
HERI solving Recurrence Equations
The recurrence relation canbe solved by following methods =
1. Substitution method ‘Masters method.
Let us discuss methods with suitable examples -
ERET stetion netnos
The subttuton metbod i a Kind of method in which a guess forthe scluton
made .
a Sara
map oa Aas Ags 10
“Tere re two fpesf substations -
Forward substation 4 Bachrard aban
Forward substitution method -
‘This method males ase of an nia contion inthe inal em an va for the
en term is generated. Thi proces i cotiued until some formule s guid. Ths
ths nd of substuton method, we se recuence equation to generate he few terms
Conse cue rion
TinTin- D+ 0
With iil onion TO = 0, Sle it using frou stitution met
Slaton:
ue
Ty) = Tn o
Wnt hen
To) = TO+1
= 041 Ini condition
mo =1 @
Wn 2, then
‘Ta) = TH+?
12 eguton (2)
a =3 o
n= then
TO) = Ta)+3
33 equation
Ta =6 ®
By obneing above generated equations we cn derive fommla
nine}) 1? a
Te) = FOE
‘We can also denote Tin) in terms of big oh notation as follows -
Tia) = O1n*)
‘But in practice, i is fc 10 guess the patter fom forward substation Hen
‘his method isnot very often use,
‘Backward substitution
{nhs method backward values are substituted rcuively in order Yo derive so
formula
TECHNICAL PUBLCATIONS® up rat owtaan aa Anya ot Aerts 130 ht
Solution + Consider, a recurece elton
Tin) = Tin-H en :
With intial condition 1) = 0
Tin) = Tia=1-1) 440-2) .
Puting equation (2) n equation () we get
Tin) = Tia-2) 4 (0-Dn a
Let
Tin=2) = Tin-2-1) #(n-2) a
Pating equation) in equation (3) we get.
Tin) = Tin-3)+(n=2)+ (a~1)n
= T-W+ (n-ke + (neko deen
k= then
Tia) = T)+1526 0
TO) =0
Tin) = 0414240040
ine
me TF
Again we can denote Tn) in terms of big oh notation as
Tin) € O(n?)
offerte
Sole cre reiton
Solon : Let
Tin) fin- 1) +1,
By backend sbsaton
Tin-1) = Tin-2)41
To) = Tin-tyet
= Minas) +1
Geom e Sect e
=
_ TECHNICAL PuBLCATIONS® A ptt rong
as an ras Ati tat
Ti) = Tin-242
‘Again Tin -2) © Tin=2=1) +1
= Tn-3) 41
Ta) = Tin-2)42
ain-3+0+2
Tin) = Tin-3) 493
Ths) = Tn- Krk 0
ike m then equation (1) becomes
Th) = TO +n
=04n a
Ta) = 0
We can denote Tn) in terms of big oh notation as
Tia) = Ot)
Soe he lowing eure
Td) «1
Ten) = AT (8) + 1? for 2 2
Soton :
= wr(8} 2
=a 9:
= wlar( BoB
TECHNICAL PUBLICATIONS® - Anup trust fo nonsdpe&
cnanution ios te ms
a eign an dns of Matt a
soa Big = —_—
wche
oa}
ala Bec |s3c
ea 3}7
I+ ec =a }pa-01¢
ua) = 2T{ 5 poke
awe pt?
rin = a
a nmty #(@- 96
gee nemeDe e121
wu
2 bn!
9
4 Cyn ABH = nlogsa A
= nl +Cn? =
Tin) = O(n?)
(SEINE se ing rene rations
vwea(z}e 79
omwer(zke ™
—————————
TERA LBLEATING? Jn te otsTo) = at? )+ 4-000
a) = aT (ye
‘ind = vm?) Assume T(N)= 1
fwertin = 51
Wake a
1 we asume 2 =n thenequton 2) becomes
Tin) = 1/2 |e logan Denk log n
Tha) = TA) + logn
Tin) = 0 dog n) Ta=t
ERED Hastr’s Neto
We cn solve recurrence relation using a formula denoted by Masters metho.
Tin) = aT init) + Fin) where n 2d and dis some constant.
‘TECHNICAL PUBLICATIONS? - A up tina ov knowledge
napand doa Aon ba os
“hen the Mater theorem ca be sated fo efircy anayias =
Ary Oi) where d 2 Oi he eran rato hen,
1 Ty = One waeo!
Ta) = Bin og n) fase
in) = 010") wae
exw undertand th Master heorem wih some exam
ae hing rene tie Ta =A +
ten We il maps eaton with
“ra) = aT (0) +f)
ow fp) is ie Hence
ped sndb= and
Dobie 2!
Tiny = @1n*) = Oxnh*4
sok
= on?) og,
Hence tie complet i in)
Fir guck an ny kono lps vales tbe 2 flowing ble can be
semi
logy =
‘romca Puen avy tinje arin of Master or 1
avo
arin)+fn) toed
fry 0 (9! hen
oy
Ti) = 86
fe) is @ (9! Tog then
si) = @ inh¥* og?)
2
Wat 82 (0! hen
ips = (0)
(ERED i stain vr nT = +
seit
tie fa) = 20g
102, bez
bog:2 = 1
Acorn 0 case 2 given above Mast tore
fin) = @(n**2* log!'n) dekel
Then Tia) = @ (nog In
6 (n"¥2" log?)
= 8k
Tin) = @{n jog?n)
(GENIE ste ig meetin Ti) = +n?
Selon
Hae)
seta ba
is
‘Then sc as foe pvr Master tren
TECHNCHL PLBLIGAMONS?. An em ty ape
ang busts es $$ St
Sanat bat Ma ba ect
fey = on)
= onthe)
2 O(n?) Mowe pute = hen O(n? -1)- O(n) hn)
Ten Ten) = 108)
rr) = 000!)
rm) = 210")
(EERE ete ve To t+
solution
ve 029 bedand fede?
and lags? =?
“according cst 3. above Master ore
ao = 2h
fe (a?) and we ave fn) en?
sen to ave =n. We mas pot =
then Th) = Olli)
Tia) = 810°)
Mathematical Analysis of Non-Recursive Algorithms
[EIED conera Pian or Anaying te Etiency of Non Recursive grt
4. Decide the input sae based on parameter
2 deny algorithms base operation).
4. Check how many times the basic operation i exerted, Then fd whether he
‘rection of basic operation pends upon the input sien. Determine wer
rrage and est eases fr input of sie ne bask operon depend upon.
rye ase, average case a est as then that has tobe analyzed separately
4. Setup a sum for the marr of ines the basi pean execu
rules
5, Simplify the sum using standard formule and
“as metoned in the sep 5 we have to simpy the sum value wing some send
Jnmala and res
TECHUCAL PUBLICATIONS? up tt rontte
[FEED surnaton Foul and Rules
= 000)
aSe
Sach
Sass
where n and k ae some upper and lower limits
ERY Examples
‘Now we wl cas & few non recurve aigorturs, We wil find the efcency
thse agots with te hl of genera plan
1. Finding th element ith maximum ae ina gin array
‘Agonthn Max Sanaa.)
‘bien Desert: Tis algorithm sor tog the
‘maxi vale element fo the ea
‘apa aay A.
‘rosa Reses te get cloner om rey
ax rave — (0) So
fori ttn do "Searing te simu leet oe
‘
MAl> Mx, raven
‘Max value € Ali a
}
rerum Max value ]. Ma. sale hen sec nen
ke ger ae
Mathematical Analysis
Step 1: The input sie is mie total numberof elements in aay.
st sic
lp 2: The basic operation is comparison in loop fr finding larger value
pre
=< RRLCATON? fn aw moma
ampmnttmsstugeis 1 sn
sup 3 +The compan ecu on eh reption cf he np As he
cans mie ena ote 0rd a ae Oa
or Pverage ee als
sora cn b te sb of ie compra Te arth
a seach tine pes Ta es wh eh ne wae
a ee Fee fr = he be comaron ses Tere
ota ote Cin a
ia) « One compaion mae for ah ale of
sep 5: Let us simply he sum
coe)
lene e(n)|
coy n-te0tn_[tavenerse
gorithm is © (0.
“Thus te efficiency of above
rents in an aay are distinc.
22 Finding whether all the el
clement wnigueness problem
pigertena Unique Elementi.)
Magetem Desc: Ths algontim inde whether aay
Felon ae distinct 02
‘gt: Ary AIO.)
pt: Raturns Fale f lements are not stint. If
‘alte elements ofan ary are ditint then it returss
tue
fori oton- 7 do
problem is
Tay two clement ine ay ae
| sir thes return Fase indaing
Pathe aay eementsate nt
dst
for} cisttanct da
f(ANT=AN) then
return False
}
}
tum True
Mathematical Analysis
Step 1: The input sie is nie total nt
‘tp 2: The basic operation wil be comparison of two elements, This operation
the innermost operation inthe loop. Hence
smber of elements in aay A.
e(Ai=Ali) then
will be the basic operation
TECHINCAL PUBLICATINS® - An op hs knowy
asp aa art Avr
eognan fy tn
step 31 The mabe of conan made wl pend wen fe Ut uy
sg Tm con ep he sone See 8 Weed a gS ths equation we wil get
se eb operon depends upon te aPuand wot ie « Fie = oti
Pane econ ep en we gl niet mp, ——.
nay ka The wos se edo by Cre) ot thee ae two oe (oP 36/20 a? 09/2
ct wont ae rp 7
pn tre we 9 eu lene in he ae
1p helt oo dees ae eile a: 20 naam
av ot ep ene compen 8 mae or xh a | CO og eon a
en eT none ng net rh ta, ou mane min
sig spon 0 = 2 ere ean Cn) 2 coe
mL ad ia at rea
Cyn) = Ober Joop x Inne loop
panes Ferengi:
Coit = EDT
step $e we wl sept Cyoy #08 = =
Fae w-y-weet [0 tenchentent]
‘The formula for multiplication ofthe above bo marie is
orrbar=an)*Bye saab awrby + aq *dy Hara]
[amsdqo rai *uottaxby aonb sandy eaaxby]
c= [Dieeas Baemaene)
[anase3toas 4x2eSxtr6x6,
‘Nove taking (1) 8a common factor, wean wnie [This canbe obtained [2 3]
sing ora ely al
. oy’ §1- Se ID
o a pees ow he algorithm for matix mulation =
ae
tiem Destin Tis elgrthm prom tiptcation
otto square rates
‘apa: Two matrices A and B
‘ome : € matrix ontaning muiplication of A snd B
for i-O to n-i do
for -Oton-1 do
Chi i)-0
fork 0t0m- do
‘tig = Cll + ADB)
(0-2-1)
= (n-t)(n-1)
(1-2)-041=(0-1)
~~ TECHNICAL PUBLICATIONS? - Anup ht konto “raven PUBLEATIONS® Anup rot enege ettosctipring A
a ee a :
wave op trot hv seth inp ler mae — —
se eprtn en aera op a which :
San eG = A ane
Visa meat te operon Both aon and mip cout coun
nL et deo hem bon ee ae
sel Mt, ol Be Od acl
|
pg one atest oer wil be counted Hence we orsier mule, 84,
cece oa
rer cpt pet Yn HP TA op
eis: Ie ite Tee tscen Hee vo" Wil Bo copa
Fe a erin WA pd noch baton pc
Ft Hece we vlc eens fr ro
Step 4; The sum canbe denoted by Mie)
Mi
‘Outermost loop ine lop x Inermast loop (exe tion)
[For loop sing] [For loop using j
> [For lop using] (1 execution)
Mon
Mia)
Thus the sinplfled sum isn. Ths the tine comy
oom eo the time complexity of matix multipin
4. Coutng umber of bts i integer
{Reena ring
Desi: Tis gene
(ary digs em a aca ictge
"Ta decal apr
VO: Renma etel nanba of gt os the apt
ip,
TECHNICAL PUBLICATIONS? « An up thes! tr inegagecg,
athenatia! Analysis
ee
Ee cad = ee
a ta oe en ee
jessie enews wcteccaiernaarss
ste et es toue tans Nee
semen inside while lop wort get executed.
Sep 3: The vale of is haved on each repetition ofthe oop. Hence tiny of
gorithm is eal 208:
step 4: Hence otal numberof tes the while loop gets exe is
Logon >
eee tie complet for counting numberof bis of given number is og 0} The
| Jindeates loor ae of og
Decl a aloith to determine the minimum ad meximum als i a
ery Ay oy iees(Hee 1 and the eres teary med wtb distin)
Demin worse compleiy ft forth agorin.— EATTRI
Solon: Alport for Minimum and maximam value in an aay
‘Ago Min Max-lamert (a (0.2)
‘arsblom Deception Tis ego is fox ning the
yeni end masa elarecs in an aay.
{pat : An aay a of ome size 2
{oat Ponsa and maximum
eament ofan aay
Max + all
Man alo,
for(i ction) do
4 (a> Max) then
Max
teal] < Max) tan
Mina i
}
Wie (The minimum element is” Mn)
‘Write (‘The maximum elements, Max)
TECHNICAL PUBLICATIONS® - An up tr! or eweMaowmatcal Analyse
‘slp 1: The input sie in fll numberof elements in ary
Bor 11 he bute sentient compan I pf hangman
ltl yeh pane Be oe ge and en
Sup 31 Thecus 5 etd on eh pet felon
aa aun
"hp Lt C0) be te maber of tom th coop ci
LET ne Thence fai =
Co) = me canaries mae fr ech a of
sp 51a py be om
aw + 2
cw «21
r
sng fra
[tetera
= 2q-te)
pias
The wort se ceny (0)
(ERRIRTE ote tg site
Algorithm Enigma (AlO...- 1, 0... ~ 1))
frie Oton-24
fajeistion=1d0
Yau jel
ri te
end for
end for
rie hue
a agri
2 What es srt cape
12 What se open?
0 Hw may i heb petion excel?
2) Mite oe sof slain?
eee ese earn a0
> TEAL PRLEATONE- wl we
is asd foradh vane of tere sno wd 0 Bnd Best case, wort cag a?
That
ex cach ne ale fhe comparisons are made, Hence for |= to mgr
beara apt tine a :
9) te rents A ab AH est ve he mai ign en
Seti ue rece tw se
10 The compre made ig sine ch wher A) gl
Alli} i the basic operation, od
ther camping te con ie of bc opt ne wil enpte want
co ee comply.
310-1)
Thus worst cae tne complet is o(s2).
iv) The bas efficency dss quadratic
1) The sigort canbe improved further by reducing to fr loops into one Yop
and obaning the te complet is O(n)
EBD consider te tong algo
Algorithn Mystery (n)
it: A a ngaie integer $=
frie ton do
SoS eivi
return S
1) What drs hs grit compute?
A Wat 6s Esc opoaton ?
Ay How many tines the Bsc operation xcued?
fn) Wat seine clas of grt. o
TEOHUCAL PURLEATIONS® Anup tr we=
uspntfoaystaerins
sotton
1) Thi algorithm ini
lagna andthe element blo dg
toara. Talis Allied
Solution
i) The ago
4 The basic operation
executed form times
{uyThe basic operation
vf sagt near bene ie compl
plesity is © ny
4) Te asc operation i itinlizatin AU
Taye ficiency fis alr tr)
{ey Te fey 8S
owen" »
1
in. [+ =
as pn ge i ef a
| Pera oa)
1 go a cg wt of es
| ceeocrmte prs ary be
tnomatical analysis of Recursive Avortis
Ph
am
'
pring (va ON»
jocuri Algor
sn for Analying the Eiincy of
plies
,
'
1 an i? ne
ai ees Dnt
i) How may ‘the basic operation #8 executed 7 2 Wentiy algorithe’'s Base ‘operations.
sey What thee SS of this algo?
ae vi int Tf
Centr ape Sas Oa
seu Se eeto eared
= yao cast Sept facie tat
Bee en scien
ii) The basic operation is ‘computat - ae nop ee
ange
sa) Basic operation * for N times ‘recurrence ‘will use the forward and backward ‘substitution rvthod
0 roe Sieh eye ren ne me
eee omc sn ion
Bo ssl! ra gage een
eae
pare
“tae
1) ie ds te grt come
i Wh he epee 7
ti Wt eis
ee
a rion?
-eomcas PUBLEATINS® At
igo ?14
essen NOE my
ee a
Selaion ——
eat comptes su of ques of 1. mnbes. Solton
ofr i) This algorithm ilies diagonal an
4) Te basi opetin su of gus of munbers eee Sti an te ene bow dag
ceo 8) Te bs operon i inlanon A ozo Tat atic
she ete a of algo ear Bens Fe competi
CENT crete Xia
1
uP
frie DON
rin dV
PePeN;
)
Ni Ph
'
i hat i is sri compte?
fat te is pat?
1 Ho mary tn Bi pein i ete?
fn Wt heen of loan ?
Er
Seton
1) Ths grt computes sum of N number, For example i
aipsiin emp 14234445018.
1) The bas sperton i computatn of sum of N numbers
49) Baie operation calf N ies.
In) The ce cs eer bcs tine complet i ON)
SIRI costes art
Aki GUESSAIND
frieOton-1
finds
Ail = 0
9 Wat es gic?
1 ate tic pation?
10 ha ey sarin 7
hen, hy
+ "Eon ALTON ton me
i) The etic of us algorithm is
1 Ein rd ef etl ls a at a ca
2 Dag a art for hing ther al eats «pe =
Dene ts worst compet 9 dint oF
EE Mathomatcal Anal
General Pian for Analyzing the Effcloncy of Recursive Algorithms
1, Decide the input see based on parameter n
2. Ientify algorithm's basic operation).
{4 Check how many times the basic operation is executed. Then find wheter the
cexcation of basi operation depends upon the input se n. Deter wore
average and best cases for np of size nf the Bsc operation depends pon
worst case, average case and best cae then that has to be analyzed sept
Se vp he recuence elation ith some itl endian eo
‘basic operation. nk
Solve the rcurence or test determine the order of growth Whie solving the
recurrence we will use the forward and backward substitution method. And
then coreemess of formula can be proved with the help of mathematical
induction method
‘Let us analyze some recursive algorithms mathematically.
EEE] Examples
4. Computing factorial of some number
‘The factorial of some number canbe obtained by performing repested multiplication
For instance: I n=5 then
Step 1: n= 51
Step 2:41 5
TECHOICAL PUBLICATIONS® «Ae up tha fr roees andes Apts ten,
Step 331645)
Step 4221630465
Step 5: 1162636405,
Step 6: 014162630405
Step Tse 1e2-30405 asole1
‘The above mentioned steps can be writen n pseudocode form a
Agorttan Feceri (2)
‘iPoblen Desrpea: Thi abort computes nung
Iecwsive tncon
‘Map: A na negative steers
‘nOurpt: tus the factor value
tin=0)
oe
tur Fata (1) +2
Mathematical Analysis
Step 1: The factorial lgoritm work for input sien
‘Step 2; The basic operation in computing factor is multpicaton
Sep 3 The recursive function al canbe formulated as
Fin) = Finn where n> 0
‘Then the basi operation muliplatons is given as Mn), And Min is mutipa
cunt to comput factorial
os Mae) +
7 N\
Th minis Tov tel
Sngundeompoe Dips
icc 6-0
Sp 4s hte eure ten band
Ma) «Mn =3) +1
Now we wil sve recurence sing
+ Toad subtton
Mat) = MO) +1
Ma) = Min) + 121+
TOM PATON? ang te
nator
Maye t=24183
wo) +
un soon
en = a2)
«| ina, stfet = Mineate2
ein 9 61161 1a Mtn =3)63
sw Mor ove coecoesof AMS formula wg maths incon a
fallow
ow Mi) = 9 bY Sg mama nd
passat 0 fen
min) = 0
ie Mg) = 0-8
ndacion we assume Min -1) = 21 then
Min) = Min =1) 41
eaettl
ie Mi)
‘us the tie compleity of factorial functions 6m
2. Towers of Hanoi
‘he problem “Towers of Hanoi” is a classic example of recursive function, The
lem can be understood by fllowing discussion. The inital setup is a6 shown in
2 | 1
TEAL PRLEATOW?. Aan hoses ct Artis _—
—_ ‘A, Band C. The five di ~,
rare as We dibs of diye oo vo os
eo el sch ay An me _ ohn
placed oP A say, : :
fe eget dk : = .
ren foes a BE ee
pacune sb aie » : |
ea Fig 174
art ap kon ar gma BONES ANY er pe, re
tha ie dks reve oneal ne oot ¢ ans
est pen esc ouie EN THe iho egy, ge 2 fom BIA
oe sk for our comfort rove dik 1 fom C10A
aso a et smb ot te ks (FOU ai femcn
: 8 . ove
A. | _[L s 8 «
ae ak Lb =
The ston an be ated =
1 Move tp n~1 dhs from Af Busing Casati Sr een
2 Moe te ering dk hom Ato C coe he cee
5. Mone hen 1 ks from Bo C sing Aas aay. ae nee
WWecan camer ito
smote disk 1 from A toB
move dek 2 from A tC 7
reve dsk 1 from B10
Fe 178
‘Tus actually we have moved n-1 ds fom peg
4 ® c
! | l ‘Agora TOHDACE)
re t ‘
Ii ont one disk has to be move
‘(a= then
Fg. 173 h
move dsk 3 from A to 8 erta|"Tho pag moved from A t0C")
retura
move dik 1 fom C104 }
move dik 2 from C to 8 a
awe top mt is om Ato Busing ©
ovedsk 1 fom A to
TOHD1A8C)
Tn Pate Aw no note TFA PRLEATONS fo atoe
oe to C using A
_—_
‘foaviacs:
: i Analy
si ane t
poeret
ET
see eee ee nS
aan ti
saot tr ernere twp Pence mks
a a eee
ns
Se arbara terse
Tae el nH Sot
Re eeererener
a nero O23
in» ten we ned vo recursive cls plus one move, Hence
vin egin-n sft fe
poe B pos (08
ne 1 thon es
To move lagest Ato C
‘om eg
To move fn 1) diy
To move (= 1) dhs
from peg B to C
fr peg Ao Bish
My = 2Mn-3) 01 | i
Step 4: Sling recumence Min) = Min ~1) +1 using two substation methods
«+ Forward substation
—
nemant trop Ane ss etn
a) = 2M) +7
22041
wes
on Fat Mig = 1) = Min -2) +1
|
| eae
mein tiem |
62 eee
af Can be writen as
4 Min = 3) +1143
[2 Mo 3)s 7822!
2 8Min=3)47
“Avove computation suggest uo compute next computation 38
Min ayy a2et
rom this we can esablish a general formula 2s
Min)
‘Tis an als be writen 8
®
[ Mn) = 2 Minnie? =
“Tus for abaining Min) we substitute n by niin the equation 2)
tet us use mathematical induction to establish corrects of equation (2).
+ Bas
‘hs in equation (2) we can obtain Min) by substintingn = ni assume iiialy
rent
te
Mia) = 2M) +2 , A
2241 —
MQ) = 3 se ees
ug) = 2a) +1 fe Ma) = Maris 2-1 when 1 an become
= 2501
= 2 E Min= (n= 1pr2" 1-1
MQ) = 7
EM (ye a= ‘Put MQ),
sym tar
TECHVCAL PURLCATIONS®- An op vat omtage TECHCAL PLATENS? - apt ote |