You are on page 1of 25

6

the fac
Determine
made made Ofproblem 108
destina
with and subjec Determine
Minimize
Thus Also and then in

0璗
machine

Example
hasn Ans.Q.
on
Mathematical 一
such
冖in I. ions
冖 clearly,since which M.A.
a

the
」 an
冖 the The Defineassignmentproblem
one-to-one

0 a 冖 machinesto

assignment
way Ⅱ a

he
assignmen
Xi total 0
「 theassignement a

M.SC.

cons Ⅳ
X
· that
at
一 objective minimum
0 assignment formulation 」 ass
(i,j only
n)

basis.

raints l,
(Final)
冖 0
each
Ofjobsis processthe

gnment
THEASSIGNMENTPR OBLE

Ⅱ problem 」
if
x~ onejob cost
ifthe assumed
一 冖
he is problem
•"
job

… 一 to
~
00 冖

,」 CO ithjob 冖
h ofthe 0
0
problem
can
Operations

n) 冖 job machines 冖~

I. SO
akes
is

:
0 is
0
jobs.
associate
incur
maximum
g
a is and
asthe beno
problem
given

assigned Aj0b … number
a
0 following a 冖 SO a
givean
special
by assigned
~ as withcost e
profit). Research

00d 一 (" are oforigins



0 0
Letxn

one Ci l,
to 冖 hej minimize 「 、
caseexample.

mathema 000h

0 and
The , Obs
」 The
冖 、
be

h a
thejth 7) Of

mach machine, only when


fora
assignment assignment

0 冖
thehe
variable

the
processedequalnum
machine, factory

冖 transportati
ne
iCal one
ove
form 」
we 冖 machi
defined

all is and is

ha CO 一
0 0
一 0 wards 一
:nimum
the envise
=
0 冖
ther lect bi W. anner
bero he
ch ethod
rarily 」then circle
l.y


the 一 0
WS. ched. lines

the冖 一 and 一
OW one
冖 冖 Stepa
in reduced

Slep Step Step 一 0 Step OW Ans.Q. The
encircled
Sle O'S
Step
Step A NO 一 一0 「 (ii)

candidate
n encircle
A
一 until 一 each Modify

Find there 一
choose

elements
) encircled 冖
2
of
、 of very …
as Rule
Subtractthis number
Ofthe
above.
do 0 Examine
冖 一 0 omStep· so0P00
Repeatthe

冖 冖 5 ma 4 3 ) 2
(v)一


一 一 a
Hungarian
… · Draw to
·
冖 not allthethis · row Write


一 ) 一 、
convenient
Tickreduced
Tick
The Go the一
一 Ow is In CO
all

reducedthe each or on.一
f
Repeat(ii) a no touch zerothe the and a
the

and
Tick
above to
lying
Draw
smallest
we
ofthese
O,the crossed.
any 冖OW
matrix theI. assignment 一
冖 冖 step
the
next single rows 「 each
modified resulting
Modify
ㄟ ㄟ column• and rows an 、
uctu
one0
( generate elements
procedure
element 「

) 一
minimum
matrix Assignment
·A
ow
assignment.
Minimum
2 matrix. s optimum
) rowstha
columns proce itera
column have column.
will
zero
·丶
lines cross
and
rows andthe elemen and
ㄗ successively

乛M.
the
0
0 「
冖 OW matrix,
Repe in and out
matrix
ive been
is 冖 intersection then

Sc
冖 in
in co algorithm.


that
as
lure epea morewill
Obviously,
each
the having

om number forthe

一 冖
that method 冖 any h -冖

assignment

assignmentis Obviously
do follows Number
ofthe be examined.Ifthere
冖 have
冖 冖
F
matrix
Intilthe not
have ofdrawingMethod. allthe × search
by

the O'S column pass



「 )
have row row.
0 the n
less OW all until repeating
columns

by
procedure reduced
ofhorizontal non-negative
encircled have of

Other

on
determiningof
uncovered than 0 0 by
chåin O'S
… the 一 「column
minimum
「 foran
ofLines
since a any
column
and
ofthe
any the
problem

row subtracting
冖 any
Opera
minimum
enci following
the made Ofthe
O'S cell

冖 op
next
two matrix0 冖
ofticking he

With
冖 encircled 一 thus of 「 are ~
O'S in
till

reduced 冖
cled) with
0
owon
冖 imal
untilnumber 一
and/or
elements de assignmentin its
an lines. 一
elements
procedure
in OW exactly

- ·
reduced

numbe an the 冖
he morecolumn. assignment
ticked
or optimal no procedure
ofthe zero the
allthe
chosen ,

zero.
assignment 冖 vertical cells matrix reduced a
一 otherv'i-•e
Research

- covered 0f0
smallest
more
entry
one
completed.
columns.
冖 and matrix. matrix. than with

0
Oflines ofencircle Proceed on

assignment
add

not lines has 」
-haveArbitrarily
and
one
matrix,the
zero
as at its

by concidered elementin convenient

in

schedule the Toop to cross is follows least columns.
一 one 0
0
any
ticked 冖
been in in found.


cover same move
imal,
0

0 and
all
any this


one
09
텷e칔죻-
명주!94좋
8 tn

(격 는 ( 결

(고 to
0 (고 듬)
(고 兮
0
0
•爻
O 들 心0
0

런 는

(鬱
수.

CO `0 (날)
0
우n국그• (들
0
O 4` 0

(날)

(수0
을 팁:@턻蓀 ;

르등•禱론 : 0흐3
(鬱(법 크

“§!i寸肄茗誓;
.屋
(사 (0
(글)
OQ 5• CO (가) ㈜ 출
(결
(길 cn 국•
0`
CO

0 0 CD
4

&8륨%
卜0
=X

一te
n matr1X.
staae. 111

tick 4
well一
om 0 二 一 column
(v) 冖
冖 0 00一
order
一 00
ma
000g巴
一 TheToThis 0
0k0
冖 一


strow

之 一 00
C 一 ow000


00 冖
= 0 = 丶
The OW
= = ·TO一
ix
一 -0W0
)
M•A•/M•Sc•

0 乛
c2 ~一above-
minimum
、 乛 m 一
0 、
0000 0 q008
一 the 0

0
-and(ㄟ
0 0
). 冖,


the
0 mo =
0
column
一00
0 0
=
ow 4 23 0 2 the
= ,
er 一
0 h 8


the 0

00 目 一

3 80 一
0-
Matrix(a) 00P e0
一 ment
wema
minimum
the

cked 0
、 · 冖wwdo,0
o
7 0 一
一 6 0
一ment 一 had minimum
00~ 一0
,0 0
, 00 no、
the we 0
(Final)

row e ∕
3 0 」 = 一
umns00一 冖一
me 000

冖000 5 一
0
2 0
8 0
c
00 not seen
- 00 00m 一 43

一、 030
0 not meh000
and4•d
= d
in 0
:= 丶
optimality'

000 一
0 :- :um9 •23 been = have
000m60


= 冖
the 0 ,
0
冖 = 一
step
: = 一
00 =
=一 一
00 0
0 乛
0 =
=

0
= 0

0 一=
= 0 12◎
40W0
00m0 0 一
u 00
m we
000
=

7
+ 一
0 m0 乛
000 4 3232 一 ow
一~ = 乛
一 = =

3'

=
0

0
乛 :
: enc 一

0
draw
一 一
0R
+ment 0 一

3一
冖 冖
0 4 ◎ 2 一 =
00 一 e
we
s 一

0 8
0-8 2 =
0
一 一
0

0 冖
ow口 0 一

0

0
not
0

9 冖
=
0 一
0

Matrix
7◎一 一6
new me0

=
一- • ow
…一we= 0
: =
0
=丶0 the 000
•乛,-0 mo0 冖 3 2 ◎ modified 冖is 0
0 一400 : drawn

0 0
冖 冖

m =

十0 0
= 一
一 draw = 0
= 0

minimum

0= =

0

0
b) 冖
=
0 0
= new cat 一
-
000
「 冖
0
一 we as
9m
目一

0
- m
肙 0-00
◎ 5 一
8 0

00ement冖
rac
= 0
」 =
一3'
which =
0 00 ,
-0=
0
「 : =
um 一
0
•hou 一
∕ 冖 ng
m- 一 nt 00 0
u 一mn•we0 = ♂

has一
0 =
lying
一 =
0 m冖
一一 a 、
0 -0 = 0 =
0
0 0

=


? -
∕ 0 mo ; 0
= 冖 乛
= 冖
∕00、
0 0
00

0

…= 一

0m00 =
=
0
:= 0
= 0 we o w : ow
一 一00

= C,returnsin 000m0m following thejobstomachinesothatthetotalcostisminimum. M
machines.

4
一 hinQ.
the -ed
Machine Q. withNow
5 ·Assign
Ans. The
following
Since As 4 3 21 profit. 6
rupees
·A totalfollowing 5 4 3
assignment 21 ·Any
The 5
·
the 65 4 3 2 1
Job
balanced

:
company
We
the given ment co0 a
冖 highest the minimum 1 0 9 7 12Il 13A
add

educed 冖
machine

assigningith
five in0
:
0 the
a

F,
a

problem
7 has
assignment

dummy

coelement 6 3 5 A
jobs cost 2 usual 19 B
0

一 9 7
5jobs 00 0 Obs 一
ofprocessing

op
冖 to 一
3♂ procedure 00
is
opportunity sixth
0
is of the machines
,00
914124 B five "1 1315 13 B
problem


supervrsor

maximization
4=
C
4
ow 0 9 4
row

subtracting 一be 0m00
3 done.
machines 4 00b

讠 of
anyjobs

loss 129 4 2616 C


Machine

4 must

ofmaximum 一
5 The C, solving 6
0
on
wishes

一 =
)
08 9 D
machineto5 00m0
in
be
all type,
so following
the
on any 丶
the
as D an 8 9 0
Findthem

a

cost
fi
elements assignment
assign

冖 One-tO-One
0 6 E 0
machine

s二
p

0 冖 maximize
the
matrix
-~

will matrix 6 is=
~
fivejobs

m Ofthe be ~ 00
given

0
nimum

9F
job ~
SO

shows problem, basis•Assign


000
matrixreduced
as

obtained. 57 6 5 4 E the(j
" 9 F
amongsiX

0
totalcost'


getthe
bellow

om to
、the
a 一 0

1
optimum
0 一
su below00amachines.
2
冖 一
0

0 一 012 2
,4 一 「per 0n5 一m二
ma 一 冖
一 0 一 This
E 0 一
2 s 一 一
h D fo 「unit 2
in T c 「 e …

neA? 「
time
「 一 mo m table
C" ; 0 nt
h

,
「…
0 一


丆「
00n 一
gis
ofmaterialshandlinginvolvingeach

e

desirable 、
es 4 3 2 1 is
0
=
minimize
C,us
es the 0fit0fRs.
2 the 5 X 一
locations• shopwhere cd 7
3 / =
than
the冖
一一& E

op

=
Others 冖一 3

imum 1 一
0
d LOC 0 is


P15-+ ◎ 乛
0

0 havea L),assignment 0


L04assign

0
machine
4 3 二
:

005
0

ma h
the

pa

+ ,Ⅴ
4
一B 1 4
X table
25 3

」 icucould4
+ and

◎ 5
2

3
」 「
new一3 7
一5

as …
4andia be ~.;
0 machines 5 5
一 machines
hand!ino•
k
0 flow
theoptimalsolutionR
2 -h ,
0 4 50.

mact…0 because
0
The and
the )n

;一 3 from;

aila Oft
th 一

冫 一
•一冖



《一
This
'冖 totaldistancetravelled? Q. =
resutting

=
0 From
h 冖
0
= Asa)csman 一
.

0
p 一
0 ciß 匕
0 the

3
tr
nmet

- usual
一 一



0、
optimum ~
8 「 「


m·冖
F000m 2
lare

20
8 ;一 proced : 04
8 23 一
= 一
msa
assignment
0
乛 一
5 27 二
;
=
0 乛


·一 36 29 乛
0 乛 冫
x)sho uld b eattached to these

~
nmerd 26 31
=
0
,
so

:

5
shown
= :
=
edalgorithm. an 2
=
below

0 assismnent
6 6
=

=
0


0

氵 「
5
,=



亖 3

“ 00 冖
m 1 ·

, ; 一
0
乛 to 冖
es unless asitgives
zero
consider

bring

C
We
assignment on. NOWcity with


Thistable

start botyh 2
intoThe 5 distance

next 3
we
and
withthe the 」
3
2 一
the

4 1

city
in minimum
4
does

6 5 3

solution.
4

to
6
according
cases

he 4 ◎ 81
6
element 4 4
not
find
are
00 13 8 6 1 13
2
separately


provide

一 ×
processed.
But冖

the


8 × 2
non-zero)
一冖
l' 6 30 13 2 2
next
6 ◎ 13


30 this

×
Table
5
the
.After
× ◎ 3
the


and 10 6 8 3
untilthe

element
best

element
1,w hi l eci t y2i snot aI I ow edt 0f 0110w ci t y
solution

Table × 6 8
optimum

7 4
10
2
make
making 8 一

solution


0 7 4
00 0
8 ◎
21 10

◎ 5
1. 26 21
acceptable

in
route

× 8
2

to

亖一0 the

occurs the

14 8
assignment ~
which

matrix 8

being
- 14 2 5
4 6
travelling

assignment ◎ 2 11 8 at

8 ◎
solution
also


- 2 8 6
0

0 2 11 14 一
,
,Therefore
satisfies

in
salesman

Places•
is

we一-cell
attained,

-00 We

this problem,

we
=一
一 n
-一
Shali extra

:0 0 乞 一
= 一
0 40~
… (a)
一 consider feasible
•0 = imes corresponding
二 00
crewofflights Crews
Flight
us :
一 vnA.M. crew
Other

0 一 0V0
resulting

d一0 4 3 2
between Ans. 4 3 2 1
一一NOW
we
一 一
0 :一
lythen Since Foreach
No.
will
must Q. The Again,if assignment

0 me quarter
The
SO

obtain 一 0
一 Similarly, F 一
p T we
be 9.

(a) =
0 ♂
feasible

一 0 be冖 一 8

0 0 6 7
1:30P.M.
selected

have An

=00a 00n 8 … on
he based

other …
30 OOA.

0-COTable
一 一 to
Delhi-Jaipur

crewho
一0
minimise

whereas Based A P.M. A.M. pa is wethis


set-up
45- 乛
pairalso air-line
Table 一
Delhi
a

一 一 02
a available

- we 一 M. can
solution

m
minimum
makeelements
atthe

and一
ruc 一 to in fl 冖
" 一
we
一 一 feasible
must
3 in 一 ; 0 一
C04
5 be M.A./M.Sc.

2 : 一
08 一
of一
03 Delhi 一
:00 二 5 00 1 assignmentin」
layovertimeaway


5 made
that
other
,The 0-一 and minutes) 7 2 9
city

mention

pur 00
have
0
,30•
according
冖 00AA「
will

fl the 一 in
taole 冖
e, 三一 0 5 termsso forthis6

000 山
a operates


8•00A.M.

一 三一
)

elements 08
layovertimes when P.M.
OP.M. 、
6 inthe
·M ;
overo

0 minimum
6
0- .00 0 一 一
the

will as ruct ,
results
on
亖 0 to 0 3, first
smallest shown
= arr一
0 crew 3 「 is solution

A.M. equivalentto 3

this zethe2
seven Final)

~ layovertime
5
be Flight
indicate + 、
hours

0 os ,
in
24 一 4 · cell 少 「ow
432 一by hours b tablesthatshow
from
layover
2
where

inm route
103 18 thesmaller

4 and
101
Hence are
marked
days

layo =flight No. Operations


2 冖
between

Delhi 0
home.

e ly
that 少
9 10
= comes
2
the

~ 冖 ~ 一
encircled


a
冖 一
0
0

一 一 b)
we 96 一乛

unit. 5 8 8
week
1.the6一 2 third
冖 no. ~
一2
the

crew
0
*)times …
30 30A.…
… OOA.
Based
0 ot units) 0 instead
out best
Jaipur-Delhi
For
h00 layover: flights.

目 9.15 一
… Depart

p ur
OONoon

de 、
r ab 一
= PM. I.
00l,the
crew 02
0
should has

一 M M.
column
· layovertimes
645P
anygiven

between programme
0 A.M.
in Research

0 2 一 一
0 0 Forsimplicity
between 0
- 一
一 一 of Table
0 03 Jaipur=
Obtain
om
time-table
一 be
75 一
一 一 冖

,M
based
0 by0 8:00A.M.6 少
be

0 9 03

and
一 ,

,
一 flights = 0 fl
flight •15A.M. 2
based. f45 2.


pairing,
the
0 : , 一 The
flights,
0 匕 104

Arrive


一crew
一 一 一 then thus

0…mNo. e when
A.M. pairing 一
below.
Delhi.
with er一 we
40
no
co 117

一 -
the
the

-
to optimum assignment 118
be 2 1
52
The or This 91 87
101 一
hours
* *
1
·A
assignments

now
total
technique,
and
89 ·冫
(3 (3
minimum 102
72 85
becomes
35
*
30 Table Sc•
minutes.

(1)一 we 冖
Final)
(1) 3 86 75 71 103


layovertimes 21
getan
(一一(
4
assignmen

02一

0
75 Operations

*
104
02 optimum 88 34 38

)一一 一

4 冖
according




03一 03 assignment
4 “2
+
0b
)一)」
4 ◎ 12
Research

(2 (2 一
em

×
101
this ·
一 The
04一
03


assignment Table
)、一 乛 Table 8 102

Ollowing
4 × 28 8
4•
103
This


w

冖 一

he
come gives

0

50 104

0 二
一 一
0

7
0 一
00~ … fthe
CYO 0
~ =
00-0 ayo 0 art 冖

= ero-sum hat elec
a d一
= = 000 :~ 日 elieve layers

一 一
minimum
(, Q.
dvantageous

乛 0 0 eg C 0 C his 冖
nits.


mo 0 minimal 」 冖
=
The 一0 一 su ma
Similarly : -.
Naturally,the♂ - wants
The TO ion
opponent
2
刁 0 n g ~ R2, move
,W1.Definegame.
Ans. Ans.

: 乛
f er 写 0 However,
4
rite
are
game
0 一 : illustrate Of

= 乛
optimal
§ 8 ma 0
upon ∕
一 a 一 = units, ra 一
n te about

a 0 、 」 0
conservative,We

the 冖
the Where

wo 0 would

4 units.∕一 player ~ 一 may 0 冖」minimise
with

move

whether 0 ed
måximin knows
shall
optimal
冖 minimum
course m
0 player
player

一 、 0 | w of thethe
IfC 丶 po 0 =
there

bepa
00
ma ximin—mmima x

0 Cwould~ : 一
now

= 冖 一 PlayerR
which = 、 0V0
00 units. 0 either 0 following
maxrmin-minrmax
that strategies
M.A./M.Sc.

0
R

R 冖一
Ch00000 b 冖
C
一 C players
the
一 ofplayer
is

R一 strategy一 c u
while ~ explain
ofaction
0 ve 00 would

emp be一 一 gains he


would a

he 一」 一
competitlon

0 like 0 L the
while

00 0
d00- 0W400 00

kew 0一 b

0 or
strategy).
like R RI 3 going
employing
the

move and
be
x

4 lose
乛 C3 to乛 0 3 3
by
RI, 冖
minimize0 0 R,C 3 4 Player
employing

player
so-called

egyoff. uni R 0 pay
CI
when 冖
Final)
the
player 」 his ♂
00 一 一 move
because
0h008
bu 冖
一 乛0

would
乛 Fo and are employ
with

C3, = 一 R unit -6 0 criterion THEORY

冖 p
C he inIfC
move
his
C2

player which
= R in
~ 一 一
0 Y0
0h00-0-
=


: in -0
diametrically
4 C ma
0
rationality Operations


moves•
players. maximin-minimax

any0 -一 - mhe冖 be
= 一
would %.
0000d such C3
his

冖 冖
R 一00 on.

-RI.R 0 0000- c, RI.e 一 RI'
0 0 二 ~ 0一 0
00
x 」 strategy

fo let
be 0 一0 = 0 m
~
-0 R Similarly,
00 his 一 :g 冖
0 一 Thus 「 us and We
冖 cannot
R C2 一
乛 nin
“ 一
=



0
口 ifRway、
op
player there

0
ca=when 、 d00
tha ra - 0000 consider
= 0 he0m0 一
乛 一
m assume
selectthat
0
亖may ~~ 0 ~
m u
RI,

that
OF Research

wo R一 、 0一 ~
f),t
冖 冖 -
0 edw ifRe 一 一 p 0
0Y0 R 一
0

is

0 0 、 一 0
player

will 0 … 一
game.

冖 = y 冖
0 Fo =
criterion

hagain 冖 0 0 0 冖 …
GAMEs

en = : 、 0 0 冖 ifR a
m 00-0- o:=
f]maximum 0 = p R h
40 冖 一
r ate : 一 一 、
0-一 vo 0
one •
= for 一 ra 宆0 0 0
00-0- Player
R


0 0
osmo RI 0
一 c 0
一given

0
which 0
二0 0冖 一 = wants
00G
both believes 一

fo 0 冖00
mo 00-his冖
forthe
e
q, Rs0
0 =乛 ~
0
119
on C
冖c R0 to the

1

payments bo them
a
markthem[tl.
payoffmatrrx by maximum game,
elementofeach 「
straidentify
120 observe of
'minimax players 0
aye
value

m
we

C2
the 、
、 Q.theStep 3 …
n 」
actually

c) This The the 、 both


· Q. Thus 一
S Step Ans. Ans.Q.3. ofthe W 冖 冖 In worst
is
can P00 "
Step
a) the
、 一 For
should5
that
4 hen The elementThe The 、 corresponding
such
based M.AJM
· gives
loss'
·Find we 「 minimum
0h00
Player



3.Ifthere 2 一
,
We A Define game.
mimmax
maximin
eade followed
results.
esulting
the
y game rowofthepayoffmatrix.
,

on Selectthe
Selectthe
maximum
minimum cases,
is see given
Pa
be 0 follow
may
rule the 「
called upon
thatthe 、

「 乛 口 Ofeach

column 一 Sc.(Final)
rom ha
made
Selected now
to
is fair 一
observe
whe The in Ofthe
payoffmatrix,
can the
the

appears said the


value

value
冖 minimum determine
game. assured
strategies
the

elementis greatest
summarize P0陽
following 冖
e 'maximin
maximin-minimax
selection
minimax
in e choo
colun

ma
maxmin
the minimax maximum
fair thatwhile
accordance
二 an mt
(ima…ble p
maximin

B
n
game 丶
enby
Operations

一 element element
are
element ofthe

一 一a 2 一 0 ofmaximin
vectorThisgives
imin 0
simple

payoff' the
value=4

saddle
value
PlayerA
Player Player PlayerA
PlayerPlayerB
saddle the
determining
losses
if m 0f4 minimax
only•The 一
payoffmatrix. value
ra
withthe
procedure value
the 一
c"
Ofeach Ofthe

Payments一 in
procedure
Ofeach

egpoint point. the row theis


is
cr units. and
the 、
"

payspays payspays
payspays 一 " 一 (4 the
minimaxvalue.
called
~ and Research
payoffmatrix 冖 v "
4 一 column
player 0
the 、
equal game.
、 5 play
R.s. Rs. IRS. Rs. IQs. Re.
from
selection Ofthe
rulesofthegamestate ow
column
oflocating
(say),
and 4

9 、
~
~
on minimax
ate
) m
the ·
(though
the the
to Ofthe
m are maximin In which
6 2 4 2 6 1 be
to {Al'
payo then player
~ strategies,
the
the maximin for corresponding

made
Ofthe
payoffmatrixand minimax havec
given
playerB
player playerA player Ofs
playerA
playerA
minimax
present player
strategies
guarantees
v
unknowingly)

the or
marked is are
ma 丶
or by
payoffmatrix
minimax
saddle nothing
value
value 、 the
given C
,
eg 一 Ⅱ value
example,


es only'冖 (4 minimtm 0 5
e mintmax
the 「
by 、
*l and th …
ther

and
point
but
given
, values,
:

Ofthe
while

mark
the

2 the
best
wo Cl
0 ).
we
and
冖 Of
0
game
一 smallest therow marked
playerA一
lumm
•ting
yoffofeachmaximum
lutions. column play
Q.冖
minima

NOand一冖
冖 一 Obviously
the Ans. Determine
iiD best v
6. 冖一
Thus
Ans.
the ClearW'
the
"
… 2
[tJ
component
value
The e 、
h
minimumFind
Mark … maxima
Since
represent
the vecto We
the
"
)
column.

ra 2
'0
ra
v also optimum

、 0 the v strategy the the


The


一 0 vector marked
00

pn matrixlargest
0
minimum
PlayerA 、 .

number m0
is
80h
The column
」0 PlayerAthe
0 obtained

40u M-AJMSc.
0 beststrategy represents maximum

rix0
solution

smallest
乛 0 column 、 e
game playerAhas
component 冖
is 」
rules
000h
maxima 0 0
乛 game
player obtained represent
2 1
yo 乛 a a by

21
saddle 23d3Pb7
with dd3 wo 2 of
to
row.
componen
not
payoffin

0 1
writing
•pe Rs.
the
乛 pa
in 0、 一
24
冖 冖
V006 0
冖 fair,
Ofr m

The
Final)
一 point by the )
一 000h PO
each
2 2
game Player
Player

, on 一 writing
The …一
3 2 represent en

0 3 22
the m
minimum
冖 冖 一 、 player. each

、m00
p
0f0 一 ow at
- ge冖
4
0

minimum Operations
ow
obtained and po given
value).
B
the

d
、 0 13 旨 0
column

ep
00m the
minima
he the 冖

esen m一
二 and
一 !
on 冖 maxrmurn
payoffin following
game
g
paydffmatrix
00
by

by onent
maximum

2
element ofthe

the writing000 ; 、

0

1)
theoptimum

一,|
…y
Research

0冖 0 pond and each


mimimax
乛 d2 pavo
payoffmatrix.

、冖 element
ofeach
the冖
ep- with 0

the value)
m、

0b 一
ng
ow

08m p
一 value matrix benefitor
maximum
op 0Y0 and
value

00d

」 二 0
=
and Ofeach
row

: 一
0
一.
ma v
those

hebyand
andThe 12
of
一 「 、
0 Thus 一
w
payo
「 22
= 一
are

b000m0 the
the

0
and

"'herep1+/)2=l
ifA

solvedA
p

0

0
刁 determined

一冖一一 一

discarded.


Q.冖two
optimum

0 一
f
一 0
plays ifB
、 matrix 一
the
When moves一
Ans.

7. strategies
= solution
iii)thevalueo ftheg ameis

-冖 + + 一 0
M.A./M.Sc.(Final)


0


一 ()2 、
、 Le ;
for 0
00optimum
PI,0冖 = 二

SA 一 」
" any m
mixed

and乛
equations
and B2
1•

a by
playerA
= the
P2 be the Thus 2 BI m
mixed
then 」 and()l


strategies

× -一
r冖
strategy
a game
= =
0 2 and
()landh va
net

乛 strategy be• two 0
PlayerA


-ne P +(12=
133; is
」冖 00 一
=
0 22 El 一
expected
person 冖
0

fo given
0 0冖 Opera
El(p)

一 冖 and
」 0一冖 0冖q) 冖q)0 冖 一 p
乛 0 = ∕ fo
forA一一
LTlievaluevofthegamet0A

0 =
I
= 冖 0
0g0 一
0 " 00 gain mo

Playerzero-sum er0L4

=
Player by

p

=
0 game*
0/21<11 一
allC/1
ed a11/71

一 1 and¯一一.B 0
2
1
0 二 optimum 一 -Bl 十 -either
-、
冖0一 0 Of 一
冖 一
B 亠
2
wi
turns 冖
= + ÷
will will 」
+

二 be
422 2 422
一 … Research

h 001122112
a12t/2 一
0
ct21P2
| | and gamefo
Bl
PI 」
0 be 一
0

given
(112
丶 ~ 冖
0
= 十 ra 一
0
be
; by 十
2
一 SIB-
w 0

:0 1)2冖
eg 一 ∕000 一
2 三 133'
= " 一 」 一
0 冖
0 ,Thewe 冖
0 一 0
any ;he
e一g一一
and m 冖
=
gain
above (12
saddle0
isgiven


00十 have 0
乛 use


: = will
= 一
eq
by

0 either
0
三 一 冖
l冖a be 一
nt
=
=

0 一
0 」
0 :0 一 -
=一= 一 d &
=
ermined
use ~
0 冖 000一
0
matching 一
0
一 00
mes
the the
Clearly, Ans.乛 一
0一0
which Q. N 冖
=
ermay equa 之
system Solving

m 0= Consider 0
ow
without

by一
∕0 … 0

=一
The
0 m rn Theaboveformulae

= 00W

ion
the

strategies.
= :
0W0
0 bo player 0
00 equationspl

the

0 Matching = 二 冖
saddle

payo 0
乛 · 二
三G0 0 maverify +
q
一 system


m = ~=、 modified一
'72 v
冖 00· 一
M.A.

乛 you T
paid
he ty "
points.


Thematrix 冖
一 ~ 00 一 E2 El(p)

PI
Player
=
00
=
∕forthe
0 0 no Rs.
0 C12
toge
P2 +192

q)

0
M.Sc.

00d00 = -
一 8 乛
0 冖
一 = "
0 ,m·00、 0 冖 yields =v=
her

mum 一
0 0 m tuti
" "
U22
p 一
Non-matching

: H m 0 0 ifthetwo the with 一
H 0= 0 乛 一 :- 一
and 冖

一 乛 一
二 」 一 E2(p)yields

' |
Final)

what m p2 乛 十 十
m

∕0 00--08 =

n g
=
ein=
g 」 0
一 冖
=
0 022 a•2 冖
2 42 冖
l)
2
—a21

= T would 0
pla 0 coinsturn
= ql o w
=|

gives 一 give

strategy 二 =
一 」
(12and
|
0
,

Operations

any 一
、 一
0 :- g
b g,0 冖 冖
一 一
0 ma 一 va 2 2
Player

、b0Y00 0
-givenpaid
一 一 00一 十 +am)
4 - 0 0
=
both 冖
一02 22
= =
inRs.= 0
00 0 ofthe
0 一
一 一

m一
0 芞 乛
ql 」
by、or 3 heads
Research

point. ·00 一 game


0
= 0-000 … 一
=

ng 一 w
、 =and 0 0m0=
0
The ,m0:=Re.
0m0
ly
乛 -0

obtained.
0

0 pla 0 0
冖 、
p
0
乛一、
01 = 0
=0 0= 一
L00 丶 一
and


123
willrepresentthe to where mixed
[3n and player given and 124

.Ifplayer
playe respectively.

Ans.
for

C p'
PureThus、
strategy Q. Clearly are Thus,
by The non-ma


0 "
Le9.
M.A./M.Sc.

moveofplayer 一
corresponding
, Write
given expec


can
0-00
the

playerBwould

emp for we

Ching
be
0 We
by optimum 冖
ed
一 汃
represented
0 、
冖 method
would


espective
value

be assume
=
Player

his 一

given
P era like
S
non-match
(Final)

、 一 ofgraphical
B 6
Ofthe
ayer
mixed

pure
being by thatthe 丶 V
matc by

like d

expected

0
to

40
I
by stra be
game

mew 8
strategies Operations

丶 the
ent 冖 +
the the
2

0
employ H
2
purestraighprobability H

4/15

|冖
game 4/15
non-matching 冖
42 solution

payoffto corresponding
Expected
In(p)= Il(p)=
12(p)=

m0V00

0
Player |3 fo

2 422

0 T T
,3
:

: 冖
冖 "
hat =
matching
/15 Research
linel' 0

ow
/15
0


:
pure Of 「
|
2""' 2
B一x
Player

payoffto
冖 冖
0
冖 0
all
(112-6122)!)+


atn-a2n)P+C12n
~ player.

P ~
employing 0、
have
4

strategy -(121)1)+
一 n)'then games. player
0

1 the
0

playerA
a
0
m
saddle : above
02n

and

the 0 his


(122 (121
乛 0XP00
Ollowing

∕0 non-matching
strategies)
pure 冖
point•

w
三 edp-
0

aye
一 冖
Letthe

a

= 冖
e
is

一 Off
on j= pointp 一 moves =ghe
Alternatively,
08that then
∕0 一 0
diffe ma
Step Since
the
00
乛 0 0 ∕imin
2

en 1• : Step
Ⅱ A 0 ; (PO) Then
point
Q. ep
0
The 00:」 1.
0 、 graphic
entyield lines
10. 彐 0L0nonaxis2..
2 一 Sp ond000
givesthe
+(p)
;
critical


em4
·一 、
0-0 ,
R00
·
Con 0 ♂=
then :
the :…
Solve nd = 」 汃=40

一 a 一
:
02 : =
一 the - 2 一 ~
m一 0 ruccan 」

the0 0 y en0 ma = 0
~00 en
moves

一 um ·、 乛 cut丶 thus
一 一
ra 0 021
1 AXIS 冖 、 2 0
一 、
0
、 ∕imin 一
identify
ra一 : 冖 game Ij
00 (p) 0W0
022

the nt he wodrawn
be
ra 一 一
o werenve


一 一 -ava
1= 一 40
ofB.

一 一0 一W0LOWER 冖 、
00 vertical
= 、 0 二 = 一 that ed0
00 00d~ op
the
0



~ on 00esent as ine 0
、 & crit …
P00 一
0P0
0 一0
二 MAXIMIN 一
:=一 axis, 一
follows 一

0m0 一
0 can 一 ma 、
(p) The
00一 0 payo 0 一
game 0
owerenve
一 0 :- (p) be 一 v、 00 、 乛
the
m

名 ENVELOPE
0
∕000 00 " axis mu 一
0 solvedmoves
determine ~
~ 0 n ax

3 whose0 00 1, ∕is 0-0000


= 0
丶 =
:
乛 冖
00、 一

0 、 冖 丶 000
m000
一 atthe 03
11 2 : 0 一
v e
p 0一
一0 、 ∕
: 、 40
乛 and一 =and (p)j=
ne
gives 3
;
0 冖 on pointp=
AXIS♂
、 0000 0 00m The、
、 payoff
0
冖 00g maximin

m (PO) 0

00n

2=
冖 」 乛 0 一 0
point 一
ma 一
0 0
00 axis =
the 一 一
t y0 0d 一 =
00-0
be :
一 一一 " 0
、一 0
ri

maximinand
p

0 一

nt 一 0 00(p) with
, easily. 00 乛 一 一
… ed ~
"一、

“ and and

0 0
= 00 & 一
0 0
一 -
ntersec
-
y
一 :
」 」

0 axis 40 0 一
:g 00a0d
2,.nt yo 2 … 0
一 “
0一
()O 0 0

identifies
一 一
0 at Ij

0 、 g

V 00 on 、
game. 00 0 that
一00
Ck 一 we
line 0 alj, 3 一 乛、 0
0
in represen payoffalj

= pa

0 payoff

second player

0 0
ersec hertwo
move
wherep' Letthe
冖 The 冖 The We Here beB'S (守一Ⅱ
ion冖
s ew
丶 冖
onpay一 he
employs
now
the
func
Oflines axis

onon payoffofthe =
Then pla) pa
pure

axis
Offlines "

p
av—a2i)p+a21,j=

axis 一 、 moves



ion 0

一 er 0
、 2 2 冖 冖
his
丶 |
envelopewi
2 、
~ 冖 AXIS
the 冖
12 3 4 5 6 7 8 9
0 l.Payoff play to
0
e
冖 0
pure
and (p) h~ ∕
10 11
2 the expec
13 shown
the
and
20
firs
A 1
nts
move 000汃
the 一a


gives
points
~
~ 一 row

LOWER
ENVELOPE 冖
ed
the

0
= 一 mixed

er
Ofthe aren 」
line
冖 乛 p
The

expected
payofflines

epresen
corre 丶 have
B
0


beingthe
0

the 一
~
n一0冖
with

1.2,3.Thuswehave

20')0 probability

s冖 冖
maximum

lines
Fig. 一
-
maximin ed obtained
冖 130))
ra wo

been

" Ⅱ
~ by
0 冖
ing
payoffto egies
Il(P)=—5p+7


2

byjoining
in
s
•(p the :d ra
2P
expected
C …
represen
一 ng payoff
一 ~ 冖
eg
po」thick
+
value

byjoining (p)} …
冖 9p
」 0
5
一 Ⅱ 0
0W0
player
n payo AXIS + 4
poI,
冖 冖 an ed as

line 冖 pay-off
2

he shown 2 ofplayerA 」

42
」 冖 on 3 4 5 6 7 8 9 moves

on
3. segmen
which 乛points 1 2 冖
10 11
theaxis
The axis when
0 in
00W

maximises
poin 一corresponding'
I. 0一
2
Fig.

s ax n and he
0 employing
p layer)


ABCD, 冖
I,j
Similarly,
representir
those below plays given
~
卜wh Ⅱ SA
=
when

2
…一
冖 」 冖
oft

冖乛 hi 一
pu
an


0
mo
value Thus
00
= and
, He or ~ 0
0- =
49


0 一
()0
0 一 .
000m0 = 00 冖

=
9
| 一
Of
er
Thusthe

一 0
0 0
000m一 … 0


0
一 must
:




0

0



p 0
、0
given
0
op

0 一

er ,; + & 0



ma
tr 2 冖
=
by
0-
♂ 一 +
0 3% 0
21
… 一


0
一 0 .ate ra :
3 0 = ♂ 一
、 、

0
+ 0 0 、 Ⅱ
0
'tncnsecuo n


2 4 0

专 ,0
,冖 = ~
0
49
player
二 then ow 00= 3
+ we
3 2
一-0 9
0 0
,二 …
: 0

5
一 乛
0 丶 23 一

2 82 、、
would
,0 -0m0
qo 00 」
and

and
9 3 Ⅱ
0
3
be 」

w
v= we
… =

0 7te 0b
= 0W0; ()o
=

~

PO
=
0 」
2
8 ~
0 一
49
"
9
p p 0
00
=

0 shown
wo ag and


040
ve 0
These B. Ans. Q.冖

一 一
Hence

0
目巴Fig• 」
Then 一
一 The
,Solve
S

·∕ axis Letthe
pure

0XP00
二一which
一 0XP00
PlayerA
the

move 4-0
with
solution

and冖
ed
40W0 丶
player
0 ed
pecte following
'1 0 1 2 3 0 一 payo
2 value
0 unit
AXIS
= Play
MAXIMIB 0 00Y0 Of
the

A
1 0 distance
p 0q00 mixed 2
0000 × 0
game
F

g.The
0冖 一
ons again 0 1
Player 4 game
game
LOWER

、 B ·∕ 00a 0
乛0 12(p)
A's
冖 3/1
-一 000
ENVELOPE

0
maximin

0 egy graphically
v 0
11
~ =
Line then =p
expected
" 9/1 8
" 00 0 SA 40
一 一 p = 丶
一 一 一2
0 Ij(p)0m0
一 冖 Ⅱ
value•

0 0
mo 、 2 2/1
ed 一 一
pay-Offl(p)

C
AXIS

4•-p 0
一 冖 0
Thus

0 乛冖
08m0d - p
0 、

ed 0 0 wherep'
2

weby :0 二
given
have ons by
一points
0
一 0一 乛
00ng
=

—P,
0
which00二
一 and then moves
the consider
thick and

the
employshis givesthe
; -
expected expected

e case, 冖 一 we
一 冖 冖
game optimal

、 ondAns.Q. 一
givesthe the 一

一 一 Forp
Hence
must Letthe Now, line


一 the 冖
h
12.Writemethode q012(p)+(1-qo)14(p)3
the
The Since

叮 一 weThe expected op and♂ PO


一 the"
payofflines
payoffequations.

optimum
segment)ABC.

| O,this have op
strategy
pu determining
冖 … function

tlintum 冖 一 consider
0 m
e expected

i mum
-0
34 h

nt
ghest

the

where move possible 冖


imal thatThe 0 player
Ⅹ 一
u yield
that
2 value 冖
一 on
forA
PO maximises
ra ra games strategy alone

… (p) M.A./M.Sc.
q p the …nes

gy 0 the0"
the nt ThepointB,
isthe moves ~
ofthe
against
forsolvinggraphicalsolutiono ~
2

is
0 point ~ Of
Wishes

eX9020d 2 This


冖 can
4 PO and
needsto
40m
inter-section

probability•that
0 ofthe game 冖 冖
乂 丶
given
5 一 ofintersection
R so =v=2/5

player similarly 一 =14()o ~4
point
function be 13's player 二

SB SA game
(p). n
that passing ~ Fig
payoffis
maximize
mixed 冖
row 2 " 2
~ "
Final)
一 一
*PO=2a
represents

thepoint
or playsolved 5、 0
冖 is 0
by
2
~ 、
"
be
5
strategy the
e
∕ 、 冖 一

determined through
onhis

Operations


「 「
00 er0 475
B2 3 5 Oflines lower
三 冖
、 丶
5 ·
4 一bythe 丶
5
the
the m of
c一 d 一 to
.Tha 」
)'= 丶 4 ~
0

mmum 一
nteenvelope
一 0
maximin

0 2 一
define 一
、 、
by (p)
冖 一
0 graphical
乛 ()o and 、

0 一column 1/5
m B4 ♂ e'
、 一
on
Research

一 「 一
expected

「「 Velo,
, 一
3 we x2games. a all =12()o the expected
(indicated
oflines12

·er co method


0 0

4
(p) 、
、、
叭 wo
一 player 一
」 Of

we

payoff

de 叭 " relevant
the
汀 一一
value

: 2

Obtain andby 一

n. 5
亠 B In the 129
0

3 we


「 「 ga
ョ 一
o
game会c 0

= 一
、 「
0 』
、 00n Fig. in
vertical
element 、
4. the
0
「 Q.オ
ed0
=
00
一 0
(9
30
一 0一
and

and


0 0 」
Then
player

0 : ).
which 20W 、
、 「
0
B 00
=.「
0 S

the
=
6th

一 axes
The
Showing A

S
Ans.
B's 一 then
he
3.0b20
9P0 オ
h
0 デ.A.
that0

( 一 00rows
on expected pu B solved
(9 0 )一

The 「
C play
y一 一
0
ミ 一
0 0 0 W04
the axis一 re 一
0
expected
匿 0 三 ま.
)
minimum


0 一 ョ
. moves onぎea00
respectively.

d = 0
= 一
p and
2 the ly. the
P 20the 一 the
g
「 0 = = 0
point 一 「 一
SC.

= 「
0 (
= 0 2
「 「
0 and 2 p
0XP020d the a
pa mixed 9
当 moves 一
0
0 0
「 一
0 ,unit

0 optimal

= (Final)

(=
minimax 一 0 「 B 0 given

【 三 オ
the
0 0
minimax

0
「 』
0 =
0 C

一 w お equations
0
Asゝ3 2
strategy
game ラ 「
0

12
(
9
) (0
辷point

0
= 「
0
A
= 「
, 「

0
0
n
(

0
お he distance
condpa
(0 payoffelement
000
p
「 6
Oblem
against

(
stra
Player
i whosea
( l
ed「aye
va

=

g
0
= 「
0P0

max 、 一
(
0
(

0
」 0
= 0

(
0
= m

0

apart.
a 日 do
24 3
A's
SB 穹
g

3 「一
0 A

0 0

0 』
(
2
(ぎ
文=
90(
9
)
— 000 0

一 譯
(
0 0 C
on 一
0
m

0
「 0

000h 一
6
(
=
(
B's ワ 0 2 16 5
pu
bo. 一
payo
0 「」
0 0 he =
0
一 n

9 4 o
了0 ー 29 一
( 』
0
・ =
n' m inimax
the0

Thusen
「 plotted
(
了 9了ーー了
9 9 「
expectednot e 2
moves 「

mpe ョ

0
=
( Re
game. 0W04

文 =
0 + 0 0
」 に
p maximum
pe ー o w 2 3q7q294q

PO
「 ed
目 0000「

0 盟

)上
the
5
minimax
3q 「
0
(=
0 コ
0
0 一
0

0 「
em on
ax一「
0「
e
line

p

+

+
6 +
5 3 pa
イ 「 三
0 a w
e
】 anysaddle 一 0 」
ラ ( ( 00
and = =(
0
( ch
( + 一 = (
=
0 0
(
0 ミ
~0
0 デ 乙 お
0
=

u nc 0

「 、

given
9

0 0

the 0 =
0 一
two 0 0P0 一 0

了 「
= 一 0
一 e (
d

0 一
・ =
ow 一 00
一 ラ ー by

.Simila
0 (一
minimax

0. 」
va 00


5=
7 一 」
= 一
0 =
0 「
0 (0
」 the0 of
q
against value =
g
「 一

「 B 's


point.

.we三0 ( 0 「
payoff' コ
point
9 」
0 一
0 0
0 .じ
mo=

0 =
P 00 theフ on会
(payoff SO. Player 0
(= 00d
「 = 0
デ一
( We
20一
= ワ me(」
0 we lines ow =

0
( 0
き一 「 which
ミき
乙 the
=
ミ 0
00
=Ans.
0.(三 (一
that
or ま
Hence

optimal
Thus
=
Ifwe
)the

0 0 一 qp
0文

4.)the
g ト0・ 2 0
value
Define optimum (

0 op
now
the 9
s. 0
』代
pu
value
00 v )+po let
ュa
m

0、
epュ

ofthe


u
(一
0
5P0
+
(一

ofp,
: ン
.A.
よと
raコ 一

ミ、
0
e

p

, gameis to (I—pオ
0 もム ANS



A

p thePO' given

0》
ト「
0
0
do一
edョ7、
まgamePO

A " " 上 " by 「
playe
Fig. 1
.


オ 5. 3
is 、
こ (F
by
p 57 、 ぎ
Theminimax

』、
0
)e an5. 0 5 775

3
Upper
envelope

「.
oppo 375 )0



be
= 0 Ⅱ

era
、 0 a

0

0 2 4 C
し ) ue.

03』

~ an

D
AXIS2

え房
00
、 ュ

0
d 0 ま
「 し

00
a 0)


0

0 ミ

0 =
0


一 0
=



= 0 9

7

0=き
0 一 】

obtained
the
the second would 132
optimum
first
Q and and
」 and Now,
Further, Ans. Q haveThe
6

column. row.

,Solvethe using …
5
、 supertor
M.A./M.Sc.

Thus 一 Use no
strategies letting

冖 we - incentive
he Hence
observe
clearthatthe
deleting
dominance
usual
followinggameusingthedominance given
strategies
we
PlayerA

method (Final)
can the
use
by a
冖 principleto
V delete
the are
| | forthe 一 、
、 d

Player
n 一
Operations
2 2 2 third d
row said
+ x + the row,the …
0

一 一 thirdcolumn Ofthe
B solvethe strategies,
| 一
| , B2

7 7 冖
7

u
冖 2

× + on column following
payoff
6) 6 6) 0f2 3 ofthe ~ Research

| 「
0 which
一 e
2 1 2 ×
2

0 reduced
getthe reduced
matrix 一
0

the

games'
ng3 are infenor
principle dominated
2 3 2 ma dominated ×
we payoff 3
2 冖rix game: ones.
easily matrix
一 domina
matrix
by A
Obtain by player
some

es the

s

Deletingreduce
reduced
bythe d

strategies.

「 columns.「
obtained.
ow affecting
also
; Ofthe
the
Ans. and(ii)(i) Thus, the
In
Il
Q.
Thus
Usedominance

thematrix

size Player
Ans.
matrix
ra
17.Two

iii) Player 5 The Now, thefifth
reduced
first row. Clearly, 一 the
ofthe the 6 reduced
we 0 The
column
(b), (a),Thus
column 一
es B solut the get…
value
aregiven must
competitorsA
must 3
ma fourth

none we given
there …
matrix
second
the 一
deleting
CompanyA
on strategies
dominates
ofthechoose
choose
pay
principle

reduced
of again payoffmatrix. to column
Ofthe
4 3 Bl 冖 dominates
0 a),
does
一 game
the column
offmatrix

4 2 4 weseethe not the s s冖 game

B3getthe exist 440


pure and
ra Of the ofthe
payoffmatrix
0 4 2 34231 find
following egy Of(b)
all second
Sc.

7
Bare
s
丶 is (c)

8 0 4 B4
payoffmam
the
ow We
、 theoptimal
1 ♂ 2 Ill … the (Final)

24 observe4
reduced 「 一一 is is
CompanvB

any
playerA 0
Ill
fi 」 competing
thus dominated one.
「 we payoffmatrix
4 2 B2 payoffmatrix
column
getthe 042 saddle
冖 e 冖
obtained「
「 )
Operations

4 24
OWS These

thatthe8 0 4
Again
and
4 and dominates

2
strategies.

0

4
point. forthe by Thus
B3 reduced can
B dominates |
4
both we
8


0 7
-
inferiorto
B4 fi
「We … for
sameproduct.
the
the
Observe
thus
Research

冖 冖
matrix play payoffmatrix
be
the

b) ow therefo first


fi
deleted
… third
is er thatthe
any dominated 3 and
冖 ·
column
a).In 冖
column. e,try the
Oftheir Their
w 一
(b) 一
the 一 … 0 and 133

d is 「
d
row 冖
b).
134
as optimum,


we
Ofthe In
Thus
Ofthe 一
However.
Q. the and Now,
are
this deleting M.A./M.Sc.

一 value strategies
using left
8 letting second
second
· reduced
Solve
with 2 1 」
we
ofgame the the
see
the and and


by forthetwo
usual
followin
matrix,
F

the n
algebraic 冖 third
third
being 0 4 column,
method
8 0 rc we
4
columns 一
convex
4 2
players B3
ws,
Operatio

v
method
" 2 see
we 4 2
0 4
x dominates

0 0
8/3. B4
2 that get
0
combination
0 and 「
the
payoffmatrix
the the
Of(b)
2
0 0
the

2 0 0 R
the
payoffmatrix.
4 game
8

convex is

9
一 PI
value the dom
on
3 3 firstrow.

92B4
whose combination

h
Ofthe Of …
34 1/3
1/3
payoffmatrix
2
game ×
2
| Thus
by

the
payoff
are
deleting first
easily
matrices, column

一 obtained the

Of
the s

value

CO

宆 ofthe
denotean Ans.
a an an
Them
an
Similarly
d
Solving Considering


Using
and
n Since d d d Let
wethese
game
value
these
for
the
obtain these
3<11|
for
p 2Pl
player
optimal
|PI—2Pl q
一qi
for

PI 一
—Pl+P2+3P5>v

+|
player
+ | +
Ofthe

|
the
value
PI
values
Simul player

the 一
十 | PI
2g2 ⅡⅡ 222
2P2
M.A./M.Sc.

4g2 2
++
22P2 7v/一
above

optimal ()2P2 十 + 2 7 冖 | +
we

2
game
ofPi
P2 | 2q3 冖
3q3 +
of

5 30.
aneous a
3P3
++
,3P3 4P2 冖
we must
eg
S P3 P2Pi q31>3
」 十+ probabili inequalities

4 3P3 汃 汃汃
(
q3 q2 一
"
as
~
>vⅣ
mus
"
=l,wegetv=15/23.

623723 一 ()2 P2 " " "
V
一 Ⅱ"
have
" Ⅱ linear
一 fo
|
ra 」
2一 9 一一 冖 、 (

丶7
46 0
丶 、 q!
2
have
020 丶
46 Players mal)
Bl g 23 冖

一一
723 一 iesin
Ⅱ = equa 丶
W46

and 2W3
2W5 ud as ()2
2
ofthe 冖
strict uperauons

B2
the
623
一一 ions

q —(11+2q2+q3=v
1q23

(j boundary

and 3

wo=
冖 in | equations,

1, q3 P3 冖 2q2
,P1+P2+P3=

1023
e
9m6
2
players = "

ms
+
respecti
3
一 2W3
conditions 3 Ofv, 4q3 Kesearcn


0
we

"

vely•
fy we get,



he Obtain Let

nequa v 一
be

一 0
column 冖
c 三
the
arc yield against
一 一 wh0%C
second
e matrix


given
: ench lontcolujnn
We 一
; Similarly, Ans. [he 三
一 Q.20. SteptheirStepwhere Od
Subtracting
一 Odd 三 Step A
)一
Step

(a) 5 三colurnn

thus
Imctit

we 一
一 冫
三 respective 一 2
C/l, a.

ow U
t9, Write nlgori(i•mforgolv ititiij/

augmentthe
get We6 ents.
Check
R
Determine
二 stjalcgics.lfnot.rite is (ithcolurnn
Augment

()btain
0 一

i"obtained


subtracting
2nd
have一 一 一
0 obtained 0
where
「 A 乛 一
二 一 ·5 二 A
, a - Let

乛 (l)tnined
column whether 一
二 rows
0
magnitudeC/ new >
the ,
the
matrix the
oddment
matrix
0 一 by 一
jeer)to kenco teo fC

2nd
Od and
from
乛 -obtained matrix
motrix ~ ~
A magnitude ~

subtracting
A
and om
method 一
m
the
columns R 」
一 from
二 口sum
byRij, corresponding
oroddments
A
R,
as

3 3 from Crom >
rd
follows method deleting
rows 1st, expressed
tosolvethe0
0 C follows
by

C 「 ~ ~ 、
55
row3-
0
anc
shown
by Imo A. isonn
- subtf'tcting

its)thcolutnn. 3rd
… A oddments
deleting

by
Hng; unc % by jnnirixjnc tltod.

3 …
Is. (afterignoring
3 from 乛
d
games in

0~ ~.

qubtracting
colurnnfiom
、 三
column
fractions ~

一co 0 三
乛 、 一 乛
· 一
Step
the
一 冖 乛 -
whosepayoffmatrices
is 一
一 w ;-1
一 l)tnntrix.
2nd
Obtaina

stand from e 3, ~

一 0
冖冖 一 「
A一
row. its


lualto ,
•the ve 一 column
2nd,ond
the isdefined


一 su
2nd g
Similarly.
2nd, grand
the
signs,
to
rows
each 一
… from
we sum ifany) VC "0

0 乛 rowson 一
OfA,
get 一
三0「 the
ow 三-

一C,
=
三 ('olutnnoddlljcnt%l
The CORow 一
0

b)
一 亠
一 C)l)tain
Here, valuc一

亖 player
an coltonn

00
player.
we 0
乛 … =m
the

are 二

0
0 冖 ow

given g
= -二 d




8
oddmct)tQ 3 25 3
三 0 c三
0
一 乛
-一 5
-given
()0 ra
PO are 5 2 60
= ~
V
1 eg …
一-

max. 冖 25 0


by8亠 丶
45 45 3 3
E 一
5 9 =
一 :
~
45 丶45 0 5 5
wo
3

22 一一
45 0
丶 一
45 丶 一 C
一 一 = 一

e
5
-一r
-given "Jot;tl%

ROW
43
2.3 oddtnentl
display
set 一
Column

and
Of We
The
a 000
eg oddments

em 00m0


0 0
value

♂ sum below

=

0
the 一 0
m
·
Ofthe

0 ow
丶3 1 2 0
game

players
∕C'
is qT ~
003 120 1 R


is: 0m0
= 0
given


1/3 冖
1/3冖


by


~
1/3, 0
3

~ ·一
0
0 ~
3
、 m00 the
+


/3
1/3)1/3)
o 丶w

× 0
=
Totals C

~
2
+ 0 00
冖 0 一
0


3


=


Row 0m0

9' 9
;
33 3 Oddments
m00
000

8

whe
concerned applicable
theNLPP Problem or then where
minimum

=I,2,...,m,and x20.

n
冖 冖
e the
g's Let
XO. subject
Ifeitherf(ll,...,xn), Ans.Q.
subject
Ans. Q.
The problem
Optimize
We

冖 {bl,
Q. (GNLPP).

withx)
Minimize
I.
be
0
2

LetzDefine

·

and
3


associate
following
the ea
The Define

epresents …
·
Usethe
b2
finding
to
the
0 classical satisfies
Ofde 一
valued be
一 general

一 a
Lagrange

constraints m 冖 bm} real


x
thefollowing M.A./M.Sc.
ermning

method
theorem


the thejth
冖 (b)
0 functions NON-LINEAR

L 一
on
Lagrange
冖 be valued
extreme
some ga
non-linear

冖 H(x) and
givesthe ∕」
十 co x
problem

ofLagrangian multiplier

X)
Ⅳ 一
the

0 一
12冖
0
c
Ⅳ"
… function

Ⅳ Ⅱ =f(x 00

」 Il-tuple
~ 乛 Of
: constants
"一variables':」
(stationary)
Lagrange
method
is (Final)

0 20 一
n called 一」
丫 involving
programming

+

equa 」
2 nx)
:
method.

Of

X22
ofnecessary
for
a


亅 …

multipliers function Opera

+ po ion. determining
General
: Ⅳsuch
variables

3X32 冖
x Then : 。
1
PROGRAMMING

•••'n

) a冖
problem-

+ condi
offsatisfring
Non-Linea-r
which xn• bl defined
0

the eq
… Finally'
orboth
the
Research

Lagrange
+
PROBLEM
0 makesz
above
… 0

& Ⅱ
extreme

constraints.
are 一
1,2 Programming

following 乛
the
,0
NLPP

6
a
method

、 宁
maximum

xo
| 一
points 一
a
一 m 一
ne


ona Ⅱ
00
139

0,] Let
is

is
0
乛 三 一 一
,一
Lagnngian 一
0
and
0b

二 Thus
following

一 Aus. 一

We

Constructthe
minima
and
e 0
0 have now M.A./M.Se.

multiplier…
We

check set
0
seen 纟
Ofthe
are
0
1 0 0
Of
ry given,

4 given
whetherthose point
necessat>' Lagrange

öL OL 冖

… 0 0
G
(Finnl)

,
G=

:=
0 problem
,x 。
| 1
Ⅱ 0
一 .
4

w 0
roo
1 00 4
30
÷
X2
3
+
condigions
一 +
function

0

era
·


T 02 0 冖
x。 ÷ 6
|X

-
given ÷
0 60 0 」 X
~ Ⅱ
128L

0 0 0 Il ,20 0 …
by
,
20 0




-a
!0
一 一

=
0
R20


global ,
000 一
·
()'
i一 ve• (I)
,
T
= 0'


:
We
=
0 二



-get

0
=
一 an conditions
:
:0
d

一 The rhe 「

0 h
=
Ans.

=
00 三 …stationary :
「 ut
theco nstraints

一 Lagrange
lere

冖 0

0
we

二 冖
二 have


0 point
0b function
M.A./M.Sc.(EinnI)

(xo.


0
C,= 一 (O) is
00

冖 一
-
3 00 0
00 0
二 given
=
02 一

0m0
=




0
~
20 =
= by
0 = 一 、
一 一
0
the
0 0


•一
o 冖
Is following
3 0 冖





;
e
-

problems

(d) (c) and (a) 、


コ(d)0 (b)
(a) 、
rhe Aい
ラub


ax2
0.6. (d) (c) (b) (a) May ア
subject A 0.5.
flaxラ

Kuhn-Tucker m .The
一 D
0 .工窪一mFO thus

yield 「
、 What to 一 コ
3
0 トヨ ula be 】 . Kuhnne
ご -ま 一
we
w

2

Kuh
暑 一 following co 《e Kuhn-Tucker 9 〒
Tucker


++ 月
X2 & の く
ゝ k
ーー 二ここ 三 十メ 聳
2 ( 0 ond CO
5)5 お
( 0
ラ )
Ⅱ X2 こ 三
(
x
) こ (こ "(x)
コ 3

00

0

0

0 0 「
0 上 ハ 」0ョ Ⅳ 0 0
日日 ハ ハ
d
三コ
ラd
0 三46 4

37
「 0. 46

+ 5. og「 の 0

cond a X2 x一
maxima
ー よ
( orOnd


x
Ⅳ ま0
「ラ 三
w
. 1
三 5 Ⅳト
.
X2+0 冐 三 ミ 0. the
0
コ 0 0. 2
) 「
0
problem

0
こ Oblem“85.
p


い 】no

u
」b 〒


e
ニc 3
0 「
programming
)ハ
0
,x
2
0 三
35 ・ 'his び ISO【 三
、 」
onditions,
tisfies does
)
2 ac
also
d08 も (c)(b)

Step
in Ans.0.8.
variables
Hence,
(
C Ca
notお e
not〔 デ Case
notCa

NOW, Let A


0
thcjth Step 0
) 04. 3. satisfy

二 0
.Vlcre.

(三
(
considered.
2.
2.
ュ the 二 二 0
W and satisfy
0
we

二 .二
)
there

). and
Const
.

non-negative
.Calgorithm
onve
d09
optimum、
0

0
But


0 (c) 3
=Oandio=().
arise 0
) こ
have 0--
str

「 andd egaand, and


(c).
一 三
uc
ニ the コ 「
not
こ こ 二 こ , +
be IV
-.
ニ 3 3
following

一 The
く 二 4 the 】 >.
2 メ
violate 上
6
Lag ith 「 he 、
(minimum)
0. 「
C「 0

0 、
quant
the
0. 「 ,こ ー 2 こ ー ー ー
二 Lag 、
ス こ 、
00n4
angianconst0
「 inequality
Orwo

n 「 一

0
In efO
「 Ⅳ
2

)
-0=
Inthiscase(a)yieldsxl 「
x
)

0
い ー
21.


「 Ⅱ any

h
e,this
hisca三 e,is
this cases:

0. 0 0;0 a コミ目
上 当 ? c.

a
三 「
a


「 2 ofthe ca ー
2 2 +ハ
(1


function
“一~ プ

. 09 s and 00
mod ミ一 お soお ace
ミ(
一 コ

(a)
discarded.
mu and 一)+2
(「 ョ

o n ( a
Kuhn-Tucke
b 一 )and
0
一 三 0
.2 一 一
a 0 (0 )y is on peaand p 0 3 一
, .2
:: 一 一 0 the 一
0

分 (b m
一 (
b

こ い ー )cr
,ミョ 婆 ハ 2
method. コ )
w given」 ) 9
ざ 一 当 .
)
2 00
,and 0
eq
=
h
い 「 0

conside b

を 0

d
Then
一 第
he上
0=NLpp 【 , 3
=

0 一
2
cond
andX2= 上 お = 一
ー by 一
the0 = 0. 三 ed.2 を
「 ー andX2 the ま

slackby . 目 is 0
ョ お a
コ -r



「 1.Th d .this
d リ Kuhn-Tucker
-
variables
O
= d け
2. 0

い 2.
一 Clcarlv 一
g 一
0
= 0k0 But



9デ
.ゴ 143
0

lltVF

'격 들. •名

크 IV 1?`|^

,71


--
tn

:굏

to 는」
는,
고i 0 누, 8
0 0 0 (글)to
0
0
0 0 9 .

수.

0 0 수 0
to .卜

0 0
0 0 0

`多
卜납

(二

(고

0
0 불 0 0 0

수. 통:
0
0 鬱4
뇨)

수 兮 0

7 0
0 0
17`
출` 0 0 0
< 00 •

CL
IA

IV
s:

tul —
and variables
•XNB arbitra given
maximization. 一
the
0 Since ForThus Now, The an Using 冖 Since
ily).
subjec
NLPP Ans. 一
00 d M.A./M.Sc.

the

~
and the XI objective
the
Thus
there 冖
to
can
Then
We
exit
e p20 will
minimum the be first
x一
function criterion,
given theare
written

|4 cons introducing
we enter
non-basic
have
p40 öX2 Of axl öf function

constraints,
two convert
(Final)

the

| a冖
expressed occurs INB XNB
constraints, in
Yl we basis.
the
s the

"0 …
0 in

f
=0 variables
compute=
0 terms standard
for
"
Öperations
minimization
in 冖
4 4x |
we

| 一 6
q30 letX3,

e the |
Ofthe
+
express
variables
are
ms 」 2-tl form
ratios'
+ 2
Of + +
R
non-basic
魚一一 andbe
INB2
the
1 2 1

3

leave 一 112 basic
X2'
the 83 objective

| weYO
and
,ch
the 4
4 22
XB
basic
write
84
b remains in in
4 terms the nc
冖 variables

1 ofnon-basi
constraints'
0


nto

chose that
2

0
0
free improved.

variable

The Clearly
1
The Thus Since For Since
We

once objective expressions


now, this
determinmg
ep
」 at

minimum 一
。 least

a
defined
will
again we
一 one

40 the enter
RIP RIP
derivatives

function have
compute ofbasic by 一
」 INB
17221
corresponds 7 leaving
the Of LA
XB NBk

=
in Ⅱ b 0 ·
offw.r.t.
the

el-ms 2 ()l
variables
2 >
{Sc.
öX2
variable, 0
partial 」
X4

the
OfXNB 」 2 3 XNB
(Final)
3 3
to cumet

2
in

」 3 are
we
derivatives 冖
erms XNB=()3
is
… 4 2 ofthe
4 3 we
find solution
2 …
have the

1 」
心一一4 3 minimum

non-basic to s
notoptimal
introduce

3 1 ratio
Rßearch
on

a and
non-bas
2
2 it
can


C 1

コ and 一


三 NOW
54

一 一 一


0 3 ン
1
the 02 こ コ ,A

0




cっ
1.

三 3 ノ
0
a
一 一
=
0
(2 (「
b
ラ 0 一
み 一
)一
4
さ 一


2 一
0
()三 9 )、 )

′一
b
イ 2 ニ
0
4
4 p






: 4 =
3 三一
0 2 一
=
0 0
ons

コ (一 (2 ス
0
コ 00 0. 8 ト
0S0

0 一
:) 3 製

NLpp 工

(2
ト ー 0
0000こ
コ 30 4
ほつ ご


)ー

-= ご

。 (第
0
= 一


u

0

0
一フ

qu2




ed
300 0.3.3. 2. 一 . 4. 3. 卩 Ans.
A 0.2. 0.
9
「 コ コ 一 ラ 一
a

a A
コ コ
latheand= =
schedule P
issuitable
「 ad.= line
maintenance
This
」 ミ
Which=0 こ (
0

CO
三 (
ToThus
pe 綉 穹 っ
3 . 挈The .D
ョu 「 ミ
e
certainty.

at OJect.
helps each ュ コ

三 ,


used u4m0
technique 0
「 ニ e used
0.0 、
0 p
一 「
ま 00ume 冐
netwo
0 mb ニ
0.

installing
me( 肴5 on「 2
一 三
m p・ 一
d
デ =
一 一 「 「 04 activ
0 、
c
ミ 一
( 一
2 a 「 etwork.
0


0
y istobem
Ymust me
comp「
0 p
「 コ 一
establishing
0
」 m 三

caidentifying
one ミ

p

ニ a 一
0
「 PERT
0

) 「
p 「
号ア
en k
Obability .A
A TurnTuGet -
e be 一
; material 「
and
was
ect be a ミ
new CS・ ミ c

task
三 titne ュp
b a




dcveloped一 0 .
time
0
Polish

made
w 一
~ 三 Which こestimateswC 一 ン
cost
a
「 「
mb )、 、イ 三
ニ =
polishedコ
0 which 一c 三 「 「
a
0

even 一
d 穹 0


三 0.


0
コ=

三一
0
he dc
「 「
0ofthe
0 「
三 0

m unce p

ec 0 一 「
0

may
he
「 一
s デ
(

=
CP 「 (F




he he 「
0


「 0

p whe
- 「
p

0
0
p 「
「 0d0
0
du
「ー一
コ ー コ
meet
calミ

o 「 コ
CPM

tnain, 「
than
0
thep 「
task.= ミ
b

PERT cp

三esentation


d w opa

0


0

ect.
a 0
conne e a

the
as

) く
e

c
such
involving 「
0

0
「 o n
allま 三
activities
)
(-)

悛 2
・ 「
寰 《
n タ
「 Ofeach
ロ コ 一
scheduled
一 e m.have Thus, 0
「 「
ect,the PERT.

pc

、 0
フ X000d
x
第 一
0
ミ o
【 p 一 a as 一

e
the 0「 三

4

) オg nd 依
0
00 一

um a

0

wi
p


0 「
e

been
d
activities
is「
equi du
activities


emphasis ac

3
=
0
30 0 =
0-ド ・
【 一 【
h co
ョ 0
ロ 一 pe
「 a

5 「
0 一

0 く
=
一 一

、 = Bo 三 balancing
フ a -0 g
三 ま ed es ま
NETWORK

p

4
交 一
o w.-
h

co
was


0

~ コ
g 「
ヨ 0
「 called
(
0 一い a

e
0

each and 代



一 デ
イ 「0 「
「 を【 っ
0
コ コ
a ed
be
no

・ ミ
be
pe needed.
given コ
Ode =

hed.
T気 【

0
一 一
d

コ0コ 0( ワ 企ec a 「

0

p
( 「
「 ロ
a
一 一
the ミ
0
ラ ョ 0 代 0
「 0
コ ご
0 「~一 0 」
00 )一
一 ゑ
つve ミ
) 、代

d 0
、 5

You might also like