Unt -I PP
(art )
|Duality tn
inear programming :
* With every lnas pongamig problem (LPP)
. P. P called
there is aluays assotated anothey
4he dual pro blerm the gtven PP
LPP foa hen. Called the onmal
oniginal
|moblem,
Primal - Dual Pairs:
Poi mal Problem Dual Problem.
Maxîmize Minimize
Subject to the constra înts : Subeet to he Constaîts :
Ax b and 70. ATw c ond wo
Maximize Mintmize z*bTuw
Subject to the consta ints: Subject to the constraints :
Az =b and n70. AwcT and w fs unTeshicted.
Minimize z=x Maxtmize z*= bw
Subeet to the constrairts: Subject to the consta ots:
Aeb ond *z0
70. ATws c and w is cuYesticte
Mancimixe /or mini míze) z=CMinimize (or Manimize)A
Subjeet to he con shraints PSubjet to the con srtrain s:
An=b and x is ATw- ct and oenmestidad
Ubresticteol
Fundamental theorem duatiy :
A necessamy and sureient conditton gor
a feasible sbluton to the primal problem,
to be opkmum is hat there
exists a Jeasite
Buch that
Soluton to tfe dual ooblem
The feastble 8oluton then Ptsetts an
he dual oblern. That is
opimum Bolutfon to
matmum fra) minimum gtw)
and
holds Lor all such pais a Xo
Mathematieal omulatan:
and epresent the number
week he oroduts A
a unit roduced per week
Then the mathematieal ormulaton
and B nespectvely.
he Eneor APP is
Marimixe ZsA02X+35 ubje ct to the (onstronts;
(Raw matenial constra int)
Ax +322 94 (Aabour hous conStraiot ).
%, 7/0 and ,70.
Sandard Primal: sntrodung slatk varfables
and a 0, the relorulated Pp la
Mamimi2e
:
Subket to he constra intt
Manimize X= Cx
x0
Subjct to the tonstraint: A-b and
Lshere c (ho, 35, D, o) ; z =[2,, , s, XJ
3
be Jbo, 96] and A-
3
Duat ! et To, , be the dual Vaab le
corTes ponding to he pofmal tonsta ints. tRen the
associated dualI o deteroine w, so
Minfmixe *= bo , +46 Wg
ubjeet o the onstaint:
3w +3w 3s
(, w, wmestfeteo)
The dual vahablesw and o, unmes tscted
ame dlomfnated by w, 0 and , 70.
Thus etimincng tfhe vedundany,, the
the esticted
Varables are and or 20.
Problens:
Pp!
1Fomuwate the dual the yolloutng
Manimize Z= 5, +372
sb. do he tonstraints 3x, + 57&15
5a, +2 2 l o
Soln:
slack vamables
SHandavd primal: áhtrolucing
he standaTd APP f !
g70 and 0,
Moxtmixe x= 5%,+3X+ [Link] +0.X
Sab. to the constaun t : 3 + 5 t 23 +O.x4 = 15
2.
Sbln.: lofte Vanables eohich is dualponoirg
Sub. torres
Sibjet Dual:
are he to problem
and Mnimize
t ae slinin
ating the let
a Minimize
dual and
te tontradrton conshants: wil w,
-to
al: and
tonshant: z= 9 be the
fe z
and wmesticted
(redundant)
are s
A e 15 onimal
4+62t 9w, thebe
ina Pp:LPP:
veoundaney, to
0,
Guation +5w2 +
10s
onstraent,
18 varhables
dual
lus 2Xg tfe O 5
restted then
les
the
the standamd omthe
PP s
Minimize
Sub. to he constrainta:
be the dual Varables
Dual: SI w, and g
prmal consta t . the dual
each primal
coYVEs ponding toto each
oblem oill be
Maximfxe
Sub. t the tonstoain
un rasdseted (redeundant)
w, ound
HBe dual oroblem îs!
redundany
Momamize z*
Sub. to the con shroin : w, A,
Dual simplex method:
A Vamant Stmplex metbod by
an
ophmum basfe feasfble soluton io
intte number steps mairtaining dual feasbilthy
as Dual
ano eomplermenlarg sfatenens
Smper metfad.
Problems:
- Use dual simplez method to solve +e CPP;
Monimize
Sub. to te tonstraint : , 472zl
Son,
Convertins he consthoung îoto
we obtain:
slack vaiables Kgo and
ntrodueio he
Love Mam z
we have:
Sub to the tonsaits! -Z, =-[Link]+0.4
-2X372+[Link] =-2
initsal basfc (nleasfble) 6oluton )9 the
An
be 29=- and x, e-2
rocokyted P ofll
-he Hevatve lual simple tables an
oleus:
Snital f4eratson:
-3
Basfe
Varfables
-2
N
Yaiable!
To ind the laving
min-,-2t
leaving vafable
To -tind the entein Vafable :
fs the enteñng vañable,
Fast iteratson!
-3
B Basic Xg
Vor.
Leawing varsabe
Vasable
’ Xg is the leaving
Enering vanfable:
max i-G
anlering varfab!e.
-3
Basfc
Vaable
-3
2
3. Vse 2.
oroblem: |LPP:
Use Sub. tfeto easible
|basie |-4 o
dual Mamimlze
o dual ophimm
Hentthe
Re
simplex constraeñb. simplen andalso
-223X,2: LP.P is
solutron apparent
metocd mettod
all
34 meached.
beenbas
+2 to X2
2. easthle
solbasfucteon
0.
to Xe; hat
X3 solve solve 706ince
2. he an'
Be al
gollowng and opimum
ollswny
(onverns tfe
mamize the dbj. dunlHon
we obaio:
mam
&ub. to snst.: X-2-R3 -1
Sntrodue og he slak varsables a o and
have he olloung LPP!
Sub. to fe conshsant:
- 3 , - 2 t g t0. + 2 -2.
An înttial baste linfeastble ) solutton to te
given P. p ib
e teractve dual simplen tables ave as
oltouos:
Snital eratfon:
1
-6 -2
Basic
Varñales
2
heaving varfa ble :
tfhe leaviog varible .
|FnBeHns \vonfable:
man
2o|=man
i6 the entesfn varsable
Pinst teratton:
Basic
Varfable
R-Neuy
-53
-to
|leaving ranfable !
nin the leavn
Varfable.
Vormable:
).
/3-d
48)
oal ieration:
-2
Basfe
Varfable
-b
Since all X- yo and also al a70,
an
ophimum baste eastble solutton has been
objained.
Hence Hhe opimun baste astble soeton
anddo
Advantages LPP:
prgramming matboals ane wed in many
4elos inddlins business and industny
* Helps fn opfimum reOurceg.
le. mani mixe proit and minimfze Costs.
Sm prove the
qualthy a decfsfons
Aimitatfons:
linear pogramntng mefho can be
applted oly i! nelatonsh+ps are (fnear.
that
* hhile soleing LPp, it fo poss?ble
we cofll get non-integral values even or thase
dects+on varfables asrch have only întegal values.
(tinear poogvomming mefBodle
are coten azsunfn atl pararmelers are knouon
and stould be ton stant. Howe ver, io real problems,
Sometmes these neither knouon
* Linasr grogrammig deal
|poblems having single obekue , choreas
eal-tile Rere many stheato ns shere
shere coe
bare to achiere
mult- objehves.
-X