wa
tenor Dnerevemene
The 3s implex method - The maximwn foe problem
Maatmun matching in Bipartite Grvecpha ~The Sable
marty ig e problem,
Lbagiive Teyerrement ;-
WE The jberabive - Improvements tech nique. involves fireling 9 eaolubon
to an opbmizabie? problem Ay generating a deynan ce ee
thle aelubena wth improving values 4 the problem's
oljedne foeten. .
yp Fach Aubsequand cin Ga) O RPE typically
involves @ Amal, Ancalired charge & the previows feosible solu
of When ne such chan dig prvoves the volur of tha objethve
fornbon, the ala ie dant fevible soluben eg
ord = fa ps)
#Smportant phme Hak car be -ehol axattly y Prerapre-
include inca paneing , woximizing he plus
inaprrovernen’ ales
ES ae rortehing She meaimum posible noe fe WA Ales
in a gr
0 Wes re blems hus :
Ffuakve ja pre ve rant
a The phm fa on aphrizakon problem , ze fir the
tha minimizes /moximizes Some value.
solubon
& An init valudon can be any und,
pdt can bk fr ther innprovee! a a segenee ee
a The -solubon vehuned when no mere Improvemen’
Powis can! ge macle.
T+ Me fee change makes an enprrovernente , bub ome
Seqyene of chanyes tan tong he impr vement# Finding an initial per be aolaton ov diffe yo mas 4
a Th mt clang clans what changes etboutel ke a lloved
in a fearible sole contig ) the. cleciaion’ manirg froters tin
Phe row changes ad already ex whng pees catia iffteat
& Lysue of Aocal vers wa plobel ene offal fe
fo[ve.
) ‘The Singh Vethod —
ae. aimglen metied pa clanie wetted jor along Tie
problem Gr wom jy _ yenusa a
Lines programy A
segyan of coe ecireme pend of he pre bile ant bbe
weqjon we im ving values of Te abjeebve pocorn
The phm of ephriaing o Dicom len area) of severed
variables du fee ae ae [near Combratsd *
manimize Cor main i2ed 0, % Toe An In
subpet fe Ay titore + Vin Ay 4 lor Zora dy fori a
yEOy---- 9 Xn BO
Guemehic. Inberpntlalion of Hasan, Programming j—
pipe nhoducing 4 ‘afluaol fot se lvrng, Ass [0r"
phos ie comider y ita 1h wilt help - fe 10
ffx fancl onmsrind pre peres of aus perm
Exnuigll Ae
omit dhe following Jinsan porrmng phn foo variables:
maximi 2d art.
sabre te xtYs & 1)
xray 2b
x20,y 20
By defrivions femible Anluson
pote 9) Bok saftper af the conabiaip
problin she reyion % fe seh oll fearsble pooh -
Bye be sketch te feasible region cy the.aAny equatin of the geneeal
a
aan b ac mt bot Lpual pr, ane ty 20, aw hue coefficient
& Sethian Ine divides Pa ph define a hough bre.
jo all the pooh & plane to poo half - planes:
; Peon % ene ef them , anebyec , whle pov
all jhe pointe tn the other ,arrhy >6-
4, potbelar, stele meets pooh defined by & equally -
gle eal pram a port on and hulow the bine ar 4 2
ant te seb 2h pein a en enters a oer
ita pink on od ea ieee eee nan en
of Since “thee pont of fhe fos ble vregion mut ahs fy all
re Conahainh hf the pm, ye fenibee yagjion b obsained by
the oferserson af these feo half ~planes ancl ths. pre prone
a He Cottesian plane clefirced hy whe sonnegabre consheurh
12 op 420 (su he
Ye
FigoFion! ple vepion of
attheb
above problem,
cot? [>
(oe) a
ty
ot Thaw , The feorrble region for the above pen. AYR chyvex
poligen with vebes 0,0), (42), Pepe
+ Ou tax ir fe feel an aphool alakon, a ye
eoritle rapier wih Ha laget voles of te offe
zosutsg,
ub the poink (ey) fo which Ra ‘eberhte PanSien 2efnrcy 5
ener. fenton cneneg
ee
chve fancheny
x
rye Sohing a Fee
imemiorol Aino
agar
probleo geome oy ;
& Sines thin Aine. deu net have commen P
feavibl poo
aegion CRG)» fhe answer bere
on tees ond, Hee ingioiely "4
fe which the ebjaeee funthor * Done ce
* tha [ine Bm 18G 7° with Pe yon 12
nea fle come
ba setbon pein’ of
4 Hur the lines an t8
Slope pz 1000? or ang othe ne
cxyhsl Jase Lines *E tha o4jebee fo"
A lines a7" i:
Ow prooblere * fe fied jes eae’
on whic
Ale vex 0?
fi ope? Sel
fa sarsyewe > 0" pate na
wily Evga 5 . sa! a
slope >! aeWe an inst ts Aine ei
ni ne tien by shy
futiovosy Called chanying " ae the line. Seb 5y
on ji 7 Ee
, oe apes aa or by shifty
Tepren fr Whe bat bine. ee pote. ”
Co i u
LE will happen ak fhe powit (3,1) with whe corntespondling
evoluc 343 + 5/=/¢-
¥ Ths means tak the ophral rolukon 1 the |inzae prrming
protein % 228 Yel, wilh Fre momimel yolue af the obgerbr
fanchon egret te It 7
haol moximize 22 ar t3¢
ebjubre fancBon an thn Aan robles, dhe devel Aue
arty oe. fo the laiges- value ef 2 would cainetae vith
eqns thak hao x ame slope 22
the boundony line 4
dhe Lwel lines
all Hee point of the [ine segment blo
& Cour epond ly
vine (3,0 and or indeding dle wobas thenweltes ,
would be epbmod solubor , gpoleling 1 flix Aame mrascmal OUR
#f fhe lg tubve chon,
& Lines ning wip He empy feoible epi on are
calleol infea ible - The kool eb infer; ble pPrr donot her
ophrood alubon
auae if ven phnia forith
Aro thas Complicakon may
fhe
pepion iy anboerded , 2 Yellewing example. larmonalrater
Excupl 8!
aH ee ver Pe wegealikes to fhe same_plm 7
aty 24 and aesye b, fa feasible pipion af te now pho
pill become un bounded, CPO).ok
LY the foul arepon of a [inser parr ing problem é
ppounded , vb obj eu ve foniion mag ex mog pop ation a
finite ophrnal value on ib:
Fot example, the problem oh manimssing 22 IX45Y peefpeut
te the tonatrainh FY Z A, ABY 2b, AZO Y2O Jor no ophmed
gol ons became Ye are pooh & feorible weapon
a) ao we wih,
a Such probleme au alld wn bouncled.
ala opbmal -cilabon 1 & finest ppenming Pom be frord
gfe. extreme pink °F ib povidle igen
Ente aed
BxtbY= 20
a, 7 3 127 2 ty [Reh | af
Simileuly, Yee 8 abve value in he y-columa the ofjerbee
cw cignals be fork Fab we to abe atreare. The volew of
fe ohpusve fuathon By intyeaving Pa velue of the See
y -cooldinat ty the nal ae ferrible SolcKon (0,0 6 od,
Tho pees 4 oe Bee eee 4 ce
sytves whee ¥z0, atuc4
Une
bith ran tho
5 gem J 4G) ataL we ine .
4 re eee a D fo a, a laigort
oes pare will Bnd ewrelves at rhe pours
Ce, it De oe adjace Jo (0, Orly b) extreme peat
cot 2 2wip> DAt8/40" fev =2 © = (Oy
TY thie ave sued nega ve andes ta the o fey
& commen, wed we fo ald eo mort negative aie,
the nega sve. nambee ith Bia (auger ab ;
ob > rule mobvared by fla obssvabrs Wot uth
chet ce gill he lacgpot inureone in sfx ehjecbve funehion’s
volun pec walt cheunge opm vastablet -vaduc
Bre evade fom ete!
”
= a+ 10 404
eRe,
Dn oun example an ireme in
at (0/04/48) changes ha vel of 2 naaesy $0u 7 OY
o &a,whik an inne th The g vee ea rey! at (0,0; 46)
changes ee fer 0 ee).
vot ja ble ny alld The eniering ven’ able
whl im column refereed bow te povet oan 93.06 Mee
the prot column by 7
A nue yous able x depart rg variable , i,
nonbuic to the nex pobLion.
. of puic youra bles wh ang done golahon
bpm, Ta 0 of eqn lig cons ean??.
ea} adjacent ex Weme pee with & Jarger value.
Fear, nel a enbilng
se ble by i danger? amouns possible {a a
yan ia 7 oe hile preening She ponnegabvitg of
a banc variable
fi obscrvaban wt Tha flews vile,
able in @ Simplex fableaus (> exch pesshiveie o the pieet win Compute the B-yako b
0M volt enby Me aig vhe pivot column,
ntaample OF foblean ©, Hoe B-mber are
Gi=4n4, 8, 24 n2,
The vow with Ha 5004 Ile,
p Brahe clebuyines Hee depats
greene Fanaa en eee ear os may be pee
KPa ew example , it variable v. We mare the row of oe
departing varia ble, calleol fhe pivet row » bys and duet it Fou:
@®
Et. alividing the
A Note tok if thee ave ne posibve enbries os thx ee
calumm, no B-rako tan be computa wkd , eohich inclicabes
poblem in unbounded and the alg. aleps
Foal, te following -stepa reed fo be taken fo framfiom o
Curren tec bloau into the next a
4 firrt , divide all the enhies of the pret you ae ee
ih enby Ha pivet column , te obbitin PF pro
& fe tableau @ ,we obra
a eG oe A,
OP ayy 3 2
& Thea, rep he eath ef the o thee rows, inclusling
ow, by tha diffocne
She eee
whue on Ke row! “entry & te five
a tableau @, tho pus
yow!- I+ ny to} + 8
oe
yew 3- 0-5) Pays g ano
Thur, the simplex mettied tr Inblean
2llewng tableau: oom i eae
10,ee
Tobloou @ vrepriesenh fe hse Penni bhe plubon (0, 22/8) with on
increaneel yale of Wo objetwe
not ophimal
cam , whith yr agra & Jo. Dr
The ext iferabon op Hole tableau ©
a d u“ v
4 Tyo tableau repost the bowie fovible solution (rt, % 2. iy
in optimal Becaue all he enkies to he a bjecdive_ mas
fablean & ave nonne gabe
«Te akimel voluc ef she objecive furcbor i eyuel fle,
the dant endy the ob Tow
Summary of the 1 2 Bh 4
ea
rT a +
ye SS cits}
t ae