Prakash M Tadvi

Asst. Prof. in Mech. Engg. Dept.,

BVM Engg. College
Learning Objectives
 Understand the principle and assumptions of Game Theory
 Know the application areas of Game Theory
 Learn types of Game Theory and various terms associated with
Game Theory
 Use of saddle point to calculate value of two person-zero sum
game with Pure Strategy
 Learn Arithmetic, Algebraic and Graphical method to solve two
person-zero sum game with Mixed Strategy
 Study Linear programming formulation of a given game
problem and its solution using simplex method.

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg College 2

Introduction
 The term ‘game’ refers to the situation of conflict and
competition in which two or more competitors have to make
decision to maximize their gain or to minimize their loss.
 The competitors are referred as ‘players’.
 Decision made by one player affects the decision made by one
or more of the remaining players and final outcome (gain or
loss to any player) depends upon the decision of all players
involved in the game.
 ‘Game Theory’ or ‘Theory of Game’ provides a series of
mathematical models that may be useful to a player to
evaluate not only all alternative course of action available to
him but also for the evaluation of the opponent’s possible
course of action in order to win the game.

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg College 3

Introduction
 This can be demonstrated with a very famous prisoner’s
dilemma game.
 In this game the two players are partners in a crime who have
been arrested by the police. Police don’t have any evidence to
prove them guilty. So, police need their confession. Each
suspect is placed in a separate cell and offered the opportunity
to confess to the crime.
 The game can be represented by following matrix:
Table – Prisoner’s Dilemma game
Not Confess Confess

Not Confess 0\0 3\0

Confess 0\3 1\1

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg College 4

Introduction
 If neither suspect confesses, both will be free. On the contrary
if both confess the crime, each will have a jail for 1 year.
 However if one suspect confess the crime he will have minor
punishment but the other will have 3 year imprisonment.
 This mathematical representation is helpful to discuss variety
of conflicting situations.
 For example instead of ‘confess/not confess’ it could be ‘set a
high price/set a low price’ for a game between two firms
competing in the same market.
 Second, from this matrix it is evident how an intelligent player
should behave. In above case, the best decision is to confess
the crime.
 If the partner in other cell is not confessing, it is possible to be
free. If the partner confesses the crime there will be prison for
1 year instead of 3 years.
P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg College 5
Game Theory
 A mathematical theory that deals with the general features of
competitive situations
 Useful when two or more individuals or organizations with
“conflicting objectives” try to make decisions
 A decision made by one affects the decision made by one or
more of the remaining competitors and final outcome
depends on the decision of all parties involved
 Applications of Game Theory:
 Two players struggling to win at chess
 Candidates fighting an election
 Two enemies planning war tactics
 Companies struggling to capture market share
 Negotiations between management and union
 Win a new contract

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg College 6

Game Theory
 Based on minimax theory (by Von Neuman): Each competitor
will act so as to minimize his maximum loss (or maximize his
minimum gain) or achieve the best out of the worst

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg

College 7
Characteristics of Game Theory
 ‘Game Theory’ is based on the principle of minimax or maximini.
This implies that each competitor will act so as to minimize the
maximum loss or maximize the minimum gain. It may be possible
that person will act where he can achieve best of the worst
condition.
 Following are the characteristics of “ Game Theory”.
1. There are finite number of players or competitors
a. If no. of participants are 2, it is “Two person game”
b. For more than 2 players, it is “n person game”
2. Each player has a list of finite number of strategy available to him.
3. Each player aware of all the strategies available to others but
doesn’t know which one of them is going to be chosen by them.
4. Each player knows finite number of possible courses of actions
available to him. The list of actions may not be the same for
others.
5. Each player knows all the possible choices available to others
(competitors) but does not know which one is going to be chosen
by them
P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg College 8
Characteristics of Game Theory
6. A play occurs when each of the player chooses one of the
courses of actions available to him. The choices are assumed to
be made simultaneously
7. Every combination of courses of action determines an
outcome which results in gains to the participant
a. Gain may be positive, negative or zero
b. Negative gain is loss
8. The gain of a participant depends not only on his decisions but
also of others
9. The gains (or payoffs) for each and every play are fixed and
specified in advance, and are known to each player. Thus, each
player knows fully the information contained in the payoff matrix
10. The players make individual decisions without direct
communication
P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg
College 9
Types of Game Theory
The game can be classified based on :
 Based on Number of Players
A game with two players is known as two-person game. If the
players are more than two the game is called n-person game.
 Based on Nature of strategies
Strategy can be called as a predetermined rule by which a player
decides his course of action from all course of action available to
him. There are two types of strategy a player can use to win a
game : Pure Strategy and Mixed Strategy
Pure Strategy : if in a game any player always selects the same
strategy then it is known as Pure Strategy. In a game of Pure
Strategy, the player knows in advance that which strategy he is
going to play and not worried about strategy adopted by
opponents.

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg

College 10
Types of Game Theory
Mixed Strategy : it is decision, in advance of all plays, to choose
a course of action for each play with some associated probability.
In a Mixed Strategy, because each player plays inconsistently
with all his available course of action opponent remains
confused about selection of his course of action.
 Based on Nature of Payoff
The quantitative outcomes resulting from a game is known as
Payoff.
Table-1 Zero-sum game Table-2 Non-Zero sum game
B B
1 2 1 2
A 1 7 -4 A 1 2, 2 3, -1

2 -3 5 2 -2, -2 5, 2

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg College 11

Types of Game Theory
Zero sum game : in such game, the gain of players that win is
exactly equal to the loss of players that lose.
For example : in above table-1, if both player plays strategy -1,
player A will win 7 units and B loses 7 units. Similarly for
strategy-1 of player A, if B plays strategy-2, he will win 4 units
and A loses 4 units.
Non Zero-sum game : in such game, all players have some gain
or loss and sum of gains and losses is not equal to zero.
For example, consider table-2, the left-hand entry in each cell is
the payoff to A and the right hand entry is the payoff to B. if
player A plays strategy-1 and B plays strategy-2, A will win 3 units
but B loses only 1 units. Hence the sum of payoff is not equal to
zero.

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg College 12

Game Models
 Based on No. of Persons:
1. Two-person game: If a game involves only two players
2. N-person game: There are more than 2 players
 Based on sum of payoffs: If the sum total of payoffs (gain or
losses) to the players is zero, it is “zero-sum game”.
Otherwise the game is “non-zero sum” game.
 Number of Strategies: A finite game has finite number of
strategies available to each player.

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg

College 13
Definitions
 Strategy: A strategy is an alternative course of action available
to a player well in advance so that he can decide which action
to choose. Strategy are of 2 types:
1. Pure strategy: If the player selects the same strategy each
time, it is pure strategy. Each player knows what the other
player is to do. Objective is to minimize losses or
maximize gain.
2. Mixed Strategy: When players use a combination of
strategies and each player is always kept guessing as to
which course of action will be selected by the opponent,
it is mixed strategy. This is a probabilistic situation.
 Optimal strategy: The course of action that leaves a player in
the most preferred position irrespective of his competitors.
Any deviation from optimal strategy would result in reduced
payoff
P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg
College 14
Definitions
 Payoff: a quantitative measure (e.g. money or market share)
of satisfaction a player gets at the end of game
 Value of the Game: Expected outcome per play when players
follow their optimal strategy.

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg

College 15
General Form: Player A’s perspective
Player A's Player B's strategies
strategyies B1 B2 B3 .. .. Bn
A1 a11 a12 a13 .. aij a1n
A2 a21 a22 a23 .. a2j a2n
A3 .. .. .. .. a3j ..
.. .. .. .. .. ..
.. .. .. .. .. ..
Am am1 am2 am3 .. aij amn
 Player A always tries to maximize his gains and is known
as maximin player
 Player B is a loser who always tries to minimize his losses,
and is known as minimax player
 A positive number in the payoff matrix represents gain to
maximizing player (maximin) and loss minimizing player
P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg
College 16
Important terms for two person zero sum game
(1) Minimax : A minimum element among the maximum elements
selected for each column is called ‘Minimax’.
(2) Maximin : A maximum elements among the minimum elements
selected for each row is called ‘Maximin’.
(3) Saddle Point : in a Game, if the maximin value is equal to minimax
value, then the game is said to have a saddle point. If the game has
a saddle point, then each player has a pure strategy.
Steps to find saddle point :
 identify the minimum cell value for each row and put it in Row
minima column on right side of matrix.
 Identify the maximum cell value for each column and put in column
maxima below the matrix.
 Identify the maximin value and minimax value.
 If both maximin and minimax values are same, the game has a
‘saddle point’.

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg College 17

Important terms for two person zero sum game
(4) Value of Game : Value of game is the average quantitative
outcome that player will win over a long number of plays.
Maximin player should use his strategy to ensure that his
average gain per play is at least equal to the value of game and
Minimax player should use his strategy to ensure that his
average loss per play is at least equal to the value of game.
(5) Fair Game : A game is said to be fair, if the maximin value and
minimax value both equals to zero.
(6) Strictly Determinable Game : a Game is strictly determinable,
if the maximin value is equal to minimax value but not zero.

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg College 18

2 person zero-sum game with saddle point
 Ex 1: Solve the game
Player B
Player A B1 B2 B3
A1 -3 -2 6
A2 2 1 4
A3 -5 -2 -4

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg

College 19
Ex. 2

Player B
Player A B1 B2 B3 B4 B5
A1 17 4 0 14 -2
A2 10 8 6 10 12
A3 2 6 4 8 14
A4 8 10 2 2 0

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg

College 20
Ex 3: Saddle Point ??

Firm Y
No Moderate Strong
Firm X advertising advertising advertising
No advertising 60 50 40
Moderate advertising 70 70 50
Strong advertising 80 60 75

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg

College 21
Game without saddle point
 If no saddle point exists, try to reduce the payoff matrix of size
of m x n to
a) 2 x 2 matrix using Principle of Dominance
b) m x 2 or 2 x n matrix and use graphical method

 Solution methods
1. Algebraic method
2. Graphical method
3. Linear programming problem
4. Iteration / approximate method

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg

College 22
Principle of Dominance
 First use maximin (minimax) principle to analyze whether
saddle point exists
 If saddle point does not exist, reduce the size of payoff matrix
using rule of dominance
i. If all the elements of row r are less than or equal to the
corresponding elements of any other row s, then rth row
dominated by sth row (Delete row r since it does not
affect column max and hence solution of the game)
ii. If all the elements of column i are greater than or equal
to the corresponding elements of any other column j,
then ith column is dominated by jth column (Delete
column i since it does not affect row min and hence
solution of the game)

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg

College 23
Ex 4: Algebraic Method
 Solve the following using Principle of Dominance and Algebraic
method

Company B
Company A B1 B2 B3 B4
A1 35 65 25 5
A2 30 20 15 0
A3 40 50 0 10
A4 55 60 10 15

Company B
Company A B3 B4
A1 25 5
A4 10 15

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg

College 24
Graphical Method
 Either of the players has only two strategies available
 Problem may originally be a 2 x n or a m x 2 game or a
problem might have been reduced to such size after applying
the dominance rule.
1. When player A has only 2 strategies to choose from and the
player B has n, the game shall be of the order 2 x n
2. When player B has only 2 strategies to choose from and the
player A has m, the game shall be of the order mx2
 Objective: to reduce a game to 2 x 2 size by identifying and
eliminating the dominated strategies, and then solve it by
1. Algebraic method (preferred) or
2. Read the game value and the optimal strategy from the
graph
P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg
College 25
Ex 5: Graphical Method
Company B
Company A B1 B2 B3 B4
A1 70 25 45 40
A2 10 60 30 50

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg

College 26
Ex 6: Graphical Method
Company Y
Company X Y1 Y2
X1 -6 7
X2 4 -5
X3 -2 5
X4 7 -6

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg

College 27
Ex 7: Graphical Method
Company B
Company A B1 B2 B3 B4
A1 19 6 7 5
A2 7 3 14 6
A3 12 8 18 4
A4 8 7 13 -1
 Principle of Dominance

Company B
Company A B2 B4
A1 6 5
A2 3 6
A3 8 4
P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg
College 28
Ex 8: m x n game: LPP
Company B
B1 B2 B3
A1 3 -4 2
Company
A A2 1 -7 -3
A3 -2 4 7
 This game has no saddle point.
 The value of the game lies between -2 and +3. It is possible
that the value of game may be negative or zero.
 Thus, a constant k (where k>=2) is added to all the elements of
pay-off matrix so that value of the game becomes positive for
applying the LPP.
 Let k = 3, then the given pay-off matrix becomes:
P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg
College 29
Ex 8: m x n game: LPP
 Let k = 3, then the given pay-off matrix becomes:

Company B
B1 B2 B3
A1 6 -1 5
Company
A A2 4 -4 0
A3 1 7 10

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg

College 30
Initial Simplex Table: Maximization
Cj y1 y2 y3 S1 S2 S3 b
S1 0 6 -1 5 1 0 0 1
S2 0 4 -4 0 0 1 0 1
S3 0 1 7 10 0 0 1 1
Cj 1 1 1 0 0 0
Zj 0 0 0 0 0 0 0
Cj-Zj 1 1 1 0 0 0

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg

College 31
Final Simplex Table
Cj y1 y2 y3 S1 S2 S3 b
y1 1 1 0 45/43 7/43 0 1/43 8/43
S2 0 0 0 40/43 -32/43 1 20/43 31/43
y2 1 0 1 55/43 -1/43 0 6/43 5/43
Cj 1 1 1 0 0 0
Zj 1 1 100/43 6/43 0 7/43 13/43
Cj-Zj 0 0 -57/43 -6/43 0 -7/43

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg

College 32
Optimal Strategies
43
Value of the game V 
13
Company B's optimal strategy
8 5
y1  y2  y3  0
43 43
43 8 8
q1  V  y1   
13 43 13
43 5 5
q2  V  y1   
13 43 13
43
q3  V  y1   0  0
13

P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg

College 33
Optimal Strategies
Company A's optimal strategy
values for x1, x2 & x3 can be obtained from the final simplex table
6 7
x1  x2  0 x3 
43
43
43 6 6
p1  V  x1   
13 43 13
43
p2  V  y1   0  0
13
43 7 7
p3  V  y1   
13 43 13
43 4
Value of the game:  3 
13 13
P M Tadvi, Asst Prof, Mech Engg Dept, BVM Engg
College 34