DEPARTMENT OF MATHEMATICS (SF)

Three Days
National Level Online Course on
Operations Research
18.05.2020 – 20.05.2020
 UNIT – V
GAME THEORY
Prepared by
Mrs. S. MAHALAKSHMI,
Assistant Professor of Mathematics (SF),
V.H.N.S.N.College (Autonomous), Virudhunagar
Tamilnadu, India.
 GAME THEORY

 The aim of the theory of games is to analyse the different

situations each player has to face and different situation he has to choose
according to those of the opponent. The term game represents a conflict
between two or more parties. Some of the recreational games such as
checkers, chess, bridge, etc., can be analysed as games of strategy.

 The application of game theory is not limited to games in

ordinary sense of it but also includes in economics, business, warefare and
social behaviour, etc. To analyse the theory of games we introduce the
terms players, strategy pay off and saddle point.
 BASIC TERMS & DEFINITIONS
 A player is called a Competitor
 In a two person game if player I have ‘m’ strategies and player II has ‘n’
strategies then the game said to be a m x n (m by n) game.
 If the number of strategies of the players is finite then the game is said to be a
finite game. If at least one of the players has infinite number of strategies, the
game is said to be infinite.
 A game with two players wherein one person’s gain is the loss of the other is
called two person zero sum game. The pay offs corresponding to various
strategies of the players are represented in a matrix called the game matrix.
 If the best strategy for each player is to play one particular strategy throughout
the game it is called pure strategy game. On the other hand if the optimal plan
for each player is to choose different strategies at different situations the game is
called a mixed strategy game.
 The optimal solution to a two person zero sum game is obtained by minimax-
maximum principle.
 Contd…
 When the maximin value = v then v is called the value of the game.
 A saddle point of a payoff matrix is that position in the payoff matrix where
. maximum of row minima coincides with the minimum of the column maxima. The
saddle point need not be unique.
 We shall denote the maximin value by 𝑣 and the minimax value by 𝑣.
 A game is said to be FAIR if 𝑣 = 𝑣 = 0.
 A game is said to be Strictly Determinable if 𝑣 = 𝑣 = v

Mixed Strategy
When maximin ≠ minimax, then pure strategy fails. Therefore each
player with certain probabilistic fixation. This type of strategy is called mixed strategy.
 Examples - Find Saddle Point
Example 1
20 15 22
Solve the following game 35 45 40
18 20 25
Contd…
GAMES WITHOUT SADDLE POINT
EXAMPLE
Mixed Strategies of player A is 𝑥1 , 𝑥2 and that of Player B is 𝑦1 , 𝑦2

𝑎22 −𝑎21 4−8 −4 4 3

Hence 𝑥1 = = = = and 𝑥2 = 1 − 𝑥1 =
𝑎11 +𝑎22 −(𝑎12 +𝑎21 ) 10−17 −7 7 7

𝑎22 −𝑎12 4−9 −5 5 2

𝑦1 = = 10−17 = −7 = 7 and 𝑦2 = 1 − 𝑦1 = 7
𝑎11 +𝑎22 −(𝑎12 +𝑎21 )

𝑎11 𝑎22 −𝑎12 𝑎21 24−72 −48 48

Value of the Game 𝑉 = = 10−17 = =
𝑎11 +𝑎22 −(𝑎12 +𝑎21 ) −7 7
 MATRIX ODDMENT METHOD FOR n x n GAMES
The method consists of the following steps
Step 1
Let A = (aij) be n x n pay off matrix. Obtain a new matrix C, whose first column is obtained
form A by subtracting 2nd column from 1st ; second column is obtained by subtracting A’s 3rd
column from 2nd and so on till the last column of A is taken care of. Thus C is a n x (n – 1 ) matrix.
Step 2
Obtain a new matrix R, from A, by subtracting its successive rows from the preceding ones, in
exactly the same manner as was done for columns is step 1. Thus R is a (n-1) x n matrix.
Step 3
Determine the magnitude of oddments corresponding to each row and each column of A. The
oddment corresponding to ithrow of A is defined as the determinant Ci where Ci is obtained from C
by deleting ithrow. Similarly oddment corresponding jth column of A = Rj, defined as determinant
where Rj is obtained from R by deleting its jth column.
 ODDMENT METHOD (Contd…)
Step 4
Write the magnitude of oddments (after ignoring negative signs, if any) against their respective
rows and columns.
Step 5
Check whether the sum of row oddments is equal to the sum of column oddments. If so, the
oddments expressed as fractions of the grand total yields the optimum strategies. If not, the
method fails.
Step 6
Calculate the expected value of the game corresponding to the optimum mixed strategy
determined above for the row player (against any move of the column player).
 DOMINANCE PROPERTY
In order to solve a game by mixed strategies, the game should be a 2 x 2 game. If it is not
a 2 x 2 game, it can be solved by dominance property. By using dominance property the problem
can be solved either by reducing it to a game with pure strategy or reducing it into 2 x 2 game
which can be solved by mixed strategies.
Rule 1
If the pay offs of the ith row are less than or equal or those in the kth row, ith row is dominated by
the kth row and hence omit the ith row.
Rule 2
If the pay offs of the ith row are less than or equal to those in the kth column, ith column is
dominated by the kth column. Then omit ith column. If a m x n can be reduced to a 2 x 2 game
using dominance then the game can be solved using the mixed strategy method.
 GRAPHICAL METHOD
If dominance property fails to solve a game problem we can try to solve it by graphical
method. But this method can be applied only to game problem of the type m x 2 or 2 x n,
(i.e) either the player A or B should have only 2 strategies.
Method of solving 2 x n game
Plot the pairs of pay off of the ‘n’ strategies of the players A and B on two vertical axes
(axis 1 and axis 2) and connect the pairs of point by straight line. Locate the highest point
on the line segments that form the lower boundary of the graph. The line that intersect at
this point identify the strategies player B should adopt in his optimum strategy.
Method of solving m x 2 game
Plot the pairs of pay off of the ‘n’ strategies of players A and B on the two vertical axes
and connect the pairs of point by straight line. Locate the lowest point on the line of segment
that form the upper boundary. The lines that intersect at this point identify the strategies of
the player A. Thereby we reduce the game into a 2 x 2 game. If graphical method also fails
to solve a game problem then it can be solved by linear programming method. We do not
discuss this method here.
 EXAMPLE: 1
Solve the following 2 x 5 game graphically
 EXAMPLE 2
Solve the following 4 x 2 game graphically: