0 ratings 0% found this document useful (0 votes) 27 views 68 pages OR
The document discusses the principles of game theory, specifically focusing on two-player zero-sum games, the maximin-minimax principle, and saddle points. It explains how players can determine optimal strategies based on their goals of maximizing gains or minimizing losses, and provides examples of game matrices to illustrate these concepts. Additionally, it touches on mixed strategies for games without saddle points and includes exercises for practice.
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
Go to previous items Go to next items
Orenations
(@) Onlytwo players participate in the game:
(0) Each player has a finite number of strategies to use,
(©) Each specific strategy results in & pay-off.
(a) Total pay-off to the two players at the end ofeach play is zero.
19.4 THE MAXIMIN-MINIMAX PRINCIPLE
bytwo players. Consider tv players 4,
This principle is used for the selection ‘of optimal strategies| oe
Aisa player who wishes to maximize his gains. while player B wishes tominimize his esses, sin
ban for player 4. the value called matimin yanee 4
ue ar
like to maximize his minimum gain, we
corresponding strategy 1s called the matimin strategy:
‘On the other hand. since player B wishes tomiimize his losses, a value called the minimax vay
{the minimum of the maximum losses is found. The corresponding strategy is called the7rinimaras,
When these to are equal (maximin value= minimax value), the corresponding strategies are called
serategies and the game is said to have a saddle point. The value ofthe game is given by the saddle,
‘The selecuon of maximin and minimax strategies by 4 and B is based upon the so-called am
minimax principle, which guarantees the best of the worst results. J
Saddle point A saddle point is a position in the pay-off matrix where, the maximum of row mi
aencide with the minimum of column maxima. The pay-off atthe saddle point is called the value
gane :
‘We shall denote the maximin value by ¥ , the minimax value ofthe game by ¥ and the value ofthe,
by.
Notes:
(i) A game is said to be fair if,
‘maximin value= minimax value =0, i.
(i A game is said to be strictly determinable if,
maximin value=minimax value#0. 1 =1= 7.
Example 19.1 Solve the game whose pay-offfatrix is given by,
Player 3B
Bh B
i ATT 34
Played 4! -4 -3
ALT Seed
‘Solution
Player B
B, By By Rowminima
meopayi sy}
wes [0 4 -3]-4
ALi 5-1} -1
Coumnmaxima | 5on
et”
or Maxims
i,
Mini(marmg, M2, 4.- Het
Maxininvate = Min 5 he 1
* Minimax value 7
ie dle point exists Thevahcethy yore
., om of the saddle point and gpg, 5 the saddle pont, which
posto jon PA), 8)
Ali 6
Player A,| -1
Al -2 4 4
puion Sener the value ofl. the pay-of matrix is given by,
: Player B
8 By By Rowminima
4
Player Ay} -1 9, -7 |-7
Al-2 4 a ]2
Cohmnmaxims ==} § 2
“The gameis strictly determinable if,
2
Example 19.3 Determine which ofthe following two-person zero-sum games are strictly determumable
and fair. Given the optimum strategy for each player in the case of strictly determinable games.
0 Player B (i) Player B
a B a B
A[-5 2]. aie
Player 4 ae “I Player 4 a -|
Maxi(minimum) = ¥ = Max(-5,~7)=-5
Mini (maximum) = 7 = Min -$,2)=—$
Solution
@ Player B
B, By Rowminima
Af-5 2]-5
Pl 1
wea 4S IE
1 4]
Column maxima -5 2
Since y = F =—5 #0, the gameisstrctly determinable. There exists a saddle point = ~S. Hence, the
Value of the game is 5, The optimal strategy isthe position of the saddle point given by, (4, B,).ast Orunations Ry
ry Player
5,8; Row minimum
Pye 4ft i]t
43 4 a
Colm manna 4.
Maxi (minivan) = y =Max(1,-3)=1.
‘Mini (maximum) = ¥ = Min (4, 1
Since } = ¥ =11#(0,thegameisstrictly determinable. Value of the gameis 1. The
is 4,3). ‘optimal strategy
Example 19.4 Solve the game whose pay-off matrix is given below.
200 5 3
Spa HAD
4302 6
5 3-4 2 6
Solution
B, B, By By Bs Rowminima
Af-2 0 0 5 3)-2
4/3 2 12 2/1
Ral oa 23 oe 6 [=a
Al S 34 2 6 ]-6
(Column maxima $3 15.6
Man (minimum) = ¥ = Max (-2,1,-4,-6)=1.
‘Mami(maxamum) = 7 = Min (5,3, 1,5, 6)=1.
Since, ¥ = ¥ = 1, there exists a saddle point. Value of the game is 1. The position of'the saddle point
is the optimal strategy and is given by, [4,, By].
EXERCISES
1. For & peme with the following pay-off matrix,
Player A
12-2
PuyrB |e 4 6
‘determine the best strategies as well es the value of the game fc A end B. Is this game ()) fair, (i)
4; game for players
{Ans. Value of game is = 2, Game is not fair, but strictly determinable]
2 Deterrine the optimal minmas strategies for each player In the folowing game,
BBB, B
A[S 207
Al 564 8
AL 402-9
TAns. g = 4, (Az, B,) Is tho optimum strategy)0"
ps WITHOUT
oat SADDLE POINTS (MIXED STRATEGIES?
95 out saddle point can be
! rg within Solved by various solution methods
A
0
ee x 2 Games without Srddie point
1s two-person 704
ya 2 r Mm game without any saddle poet hmv Se nyo maser player
cans
fe & B,
ae a
a az: }
me optimum mixed strategies,
apt ps2 =P"! Fi
wheres
te __9+q,51> elm
aireleeai cae
%
4 9 = 42
(a4) + a2) ~ (42 + 221)
‘The value of the game (j) =
5 Solve the following pay-off matrix. Also determine the optimal stravegres and vale of
Example 19.
‘the game-
B
ee
34
B
51), Lec this be,
34
Solution
ae
4{ a, a2 ] “The optimum mixed straseBies,
Alay, an
(& 3)
s-(4 8) iss ( a
“la La :
ayy - Oy 4-3 1
i tc} at
z
ss PL” Tay, Pay) (0g 220 *a-aed
poten 2 Pl3H Oremmnons
4 SR
Sy
22-4, ae
= sete;
1” Gy Fax) ~(Oi2+@y O¥4) 09 75
3_2
a7 n2t- 525
x= Usa) TD
value of game. . (5+4)-+3) 5
The optimum mixed strategies.
Example 19.6 Solve the following game and determine its value.
B
[<<]
4
44
Solution Wis clear that the pay-ofT matrix does not possess any saddle point. The players will use mixe
strategies. The optimum mixed strategy for the players are,
4 *)
Se sp, +py=1
‘ . my
B, )
Sp= nth?!
i Fe Gide
=a 4 Bt
PL Gay + ayy) (aj 4a) 4+4-(-4-4) 16 2
a.
pytl-py > 71-575
eee ce
N” Gran) (q, ay) 444-(-4-4) 162
ee
=1-9,=l--=2-.
a7 1 7-2
memmiens( f)-($ 9
The value of the game is, y= —229u~ 12421
(agg + 4)~ (a2 + a1)
(44)=[-44(-a)A ie
oe
19.7, 93 ea fa os
A ga wins nothing we ee
hha cere Spree aa Seg to
te aa aslo Pun oftaewhen tee rend
ow rae player and the value of the
104 ye pay-off matrix forthe
On aye A ispivenby
patthis be
pee optimum mined StateBis,
—a-
4 (ayo
1.3
gr h-qr ary
‘Value of the game = 0-4-5)
+0-
\
8436 Orererny
Me
The optimum mixed strategy is given by,
_ hind the value of game is -1/8.
EXERCISES
1 For a game wan the folowing payff mats, determine the Optimal stretagy and the value o hy
o OB B45 Sane.
6-3
43 3)
By One
mw 8
i 4)
2. Two players, A and B match coins. Ifthe coins match, then A wins two units of value. Wf CONS do ret mg
then B wins two units of value.
Determine the optimum strategies for the players and the value of the game.
[ane SaSp
3. Consider a ‘modified form’ of matching-based coins game problem. The matching player is pad © 600 1
the coins tum heads and € 1 if both the coins tum tails. The non-matching player is paid & 300 when te
‘coins do not match. Given the choice of being the matching or non-matching player, which one was
choose and what would be your strategy?
[Anns Snaen*( 75} Sterna “| 2 ¥) Value of game = ~+,Non - matching pin
Hint: Pay-off matrix for matching player is, Non-matching (B)
HT
uf 8 -3
Matching player (A) 7] _y
19.5.2. Graphical Method for 2 n or mx 2 Games
Consider the following 2 games.
Strategy for player B
By B,
Strategy forplayer A |» Dt Br Fe
. Pier Aan a2 On
Ay Lay) 422 + Faq
Let the mixed strategy for player A be given by,
A
S,> ‘) such that, p, +p," 1,p), p20
PL Pz :eo
eat of the PUFE HHERCS ite 1g p oa
0
wo prspuremon® been pay-off for player A would be as follows-
B, é Pected pay-off Elp)
By Berea
20)™ 012, + ay>Py
B aie
i n(P)~ 249, + dayPy
ike to
m8 wos haces cui See! is a minimum for
of player isto eiepech (Efoh=1.2.-
the straight lines, and pin sucha way that, 7a large as possible. This may PF
Bie) = 0,0," 0,"(0, 4) Py *
i225"
jective
tions of P;
‘avon the lower boundary/ofthes ns wil gi i
Il give the maxim
in the lower boundary (lower envelope) as well asthe optimum:
neat ;
ehighest Oa um valueamong the minimum
peed pay-ofls value of probability P;
of player B corresponding tothe lines th ee
2 the can
a theps in reducing thesize ofthe game 2) at pss tah he main PN
sntreat m *2 games in thesame way and get the mii
we cal
pper poundary (upper envelope).
oP
Now the two strategics
ecerrit
max point, which will bethe lowest
4g Solve the following 23 game graphically.
Player B
1,3
Player A
vet [hs a
Solution Since the problem does not possess any saddle point, let player 4 play by the mixed strategy
(* 4) sitio t-P
A Pa
sainst player B.
are etd pay-off gains B's puremove Sve
B's pure move arsexpected pay-off
Elp))
B, gyp)prvs-0)— 78
By gyppr snes -Pd 57
By ego iee PnP