Chapter 3 Game Theory

1.1 Introduction
1.2 Definitions and Theories of Game Theory
1.3 Introduction to Software
1.4 Applications
1.1 Introduction
Game Theorists who have received
the Nobel Prize
 1994 ：纳什（ John F. Nash ）、海萨尼（ John C.Harsanyi ）、泽尔腾 ( 德）（ Reinhard.Selten ）

• Non-cooperative game
• Incomplete information
• Subgame and dynamic
game
 1996 ：莫里斯（英）（ James A.Mirrlees ）和维克瑞（美）（ William Vickrey ）

• Asymmetrical game
• Application
 2001 ：阿克洛夫（ George A. Akerlof ）、斯宾塞（ A. Michael Spence ）、斯
蒂格利茨（ Joseph E. Stiglitz ）（美）

Market for Lemons and

Asymmetric Information
 2002 ：弗农史密斯（ Vernon Lomax Smith ）

Psychology into
game theory
 2005 （以色列）奥曼（ Robert J. Aumann ）、谢林（美）（ Thomas C. Schelling ）

Conflict analysis
into game theory
 2012 哈佛大学教授埃尔文 ·E· 罗斯、（ Alvin E.Roth ）及加州大学罗伊德 ·S· 沙普利
（ Lloyd S.Shapley ）

Theory of stable allocations

 2017 芝加哥大学商学院教授理查德 ·H· 塞勒 (Richard H.Thaler)

He is a theorist in behavioral finance and decision

science. In 2017, he was awarded the
Nobel Memorial Prize in Economic Sciences for his
contributions to behavioral economics.
John Forbes Nash Jr.
 Family of Game Theory
Father- von Neumann

Fraser 和
Hipel

Kilgour 和
Hipel
 Professors in Waterloo

Keith W. Hipel D. Marc Kilgour

1
 Example 1: Prisoner’s Dilemma

 Two suspects arrested for a crime

 Prisoners decide whether to confess or not to confess
 If both confess, both sentenced to 3 months of jail
 If both do not confess, then both will be sentenced to
1 month of jail
 If one confesses and the other does not, then the
confessor gets freed (0 months of jail) and the non-
confessor sentenced to 9 months of jail
 What should each prisoner do?

Game Theory Jan 07, 2009 1

 Game in Normal Form

 Normal Form representation – Payoff Matrix

Prisoner 2

Confess Not Confess

Prisoner 1 Confess -3,-3 0,-9
Not Confess -9,0 -1,-1

1. What is a solution?
2. Is it important of the value of months?

Game Theory Jan 07, 2009 1

 Game Theory
1. What is Game Theory?
2. What is a Game?
3. What are its Components?
(1) The study of mathematical models of conflict and
cooperation between intelligent rational decision-makers.
(2) The game is the fundamental model of game theory.
1. A (well-defined) set of 2 or more players
2. A set of options (or actions) for each player.
3. A set of preference relationships for each player
for each possible action tuple.

1
What is a Game?
 A game consists of
– a set of players
– a set of strategies for each player
– the payoffs to each player for every possible list
of strategy choices by the players.

1
 PREFERENCES
 Would you like to have tea or coffee to drink.
– Quantitative response: I have a utility value of 1.9673 for coffee
and 1.0000 for tea.
– Human response: I would prefer to drink tea. Thank you.
 Quantitative preferences:
– Cardinal number is assigned to each state or object (ex. Dollars or
utility value).
 Non-quantitative or relative preferences:
– One state is either more preferred, less preferred or equally
preferred to another.

1
 Preference Relations
Preference Relation expresses an individual player’s desirability of
one outcome over another (A binary relationship)
i Preference Relationship (prefers at least as much as)

a i a* a is preferred at least as much as a* by player i


i Strict Preference Relationship (prefers strictly more than)

a i a* iff a i a* but not a* i a
 

~i “Indifference” Relationship (prefers equally)

a ~ i a* iff a i a* and a* i a
 
1
 PREFERENCES

 Cardinal
 Ordinal
 Strictly Ordinal
 Unknown
 Greatly more (or less) preferred
 Transitive
 Intransitive

1
 PREFERENCES
 Ordinal: States are ranked from most to least
preferred where there can be equally preferred states.
 Strictly Ordinal: States are ranked from most to least
preferred where there are no equally preferred states.
 Transitive: If a decision maker prefers state p to q
and q to r, this implies he prefers to p to r.
 Intransitive: A decision maker prefers state p to q, q
to r, but r to p.

1
 Elmira Conflict
Elmira is located in southwestern Ontario, Canada. In 1989,
the Ontario Ministry of Environment (MoE) detected that the
underground aquifer supplying water for Elmira was
polluted by N-nitroso demethylamine (NDMA). A local
pesticide and rubber manufacturer, Uniroyal Chemical Ltd.
(UR), was identified, since the prime suspect as NDMA is a
by-product of its production line. Hence, a Control Order
was issued by MoE, requiring that UR take expensive
measures to remedy the contamination. UR immediately
appealed to repeal this control order. The Local Government
(LG)} consisting of the Regional Municipality of Waterloo
and the Township of Woolwich, sided with MoE and sought
legal advice from independent consultants on its role to
resolve this conflict. 1
Utility Functions

A mathematical description of preference relationships.

Maps action space to set of real numbers.

ui : A  R

Preference Relation then defined as

a i a* iff ui  a   ui  a 
*


a i a* iff ui  a   ui  a* 

a ~ i a iff i    i 
*
*
u a u a

2
 DRAWBACKS OF CLASSICAL
GAME THEORY
 Preferences over states for a decision maker
are represented cardinally by von Neumann-
Morgenstern utilities.

These are notoriously difficult to calibrate.

People think in terms of relative preferences.

2
 When Conflict Arises
 A strategic conflict is a situation in which two or more
decision makers (DMs) with different objectives interact.

 Assess whether it is substantive or due to

personality and temperament differences.
 If due to temperament differences, filter–relate to
the other in his/her terms.
 If conflict persists, identify and agree on the item
or issue in dispute.
 Negotiate on that item or issue.

2
 Negotiation
 Human interaction is essentially negotiation.
 Negotiation is interaction.
 We must become aware of our behavior,
preferences, and mannerisms.
 Not everything SHOULD be negotiated:
 Potential for MUTUAL BENEFIT is a prerequisite to
negotiation.

2
 Negotiation Style
 Avoiding may sustain positive outcomes but
permits escalating negative outcomes.
 Competing creates a win/lose game.
 Compromising gives up something.
 Collaborating creates win/win.

2
1.2 Definitions and Theories of Game Theory

2
GAME MODELS
 Normal or Matrix Form
 Option Form
 Graph Model

2
Chicken Game in Normal Form
PLAYER 2
Don’t Swerve Swerve
(D) (S)
1 2
Don’t Swerve (D) 1,1 4,2
DD DS
PLAYER 1

Swerve (S) 3 4
2,4 3,3
SD SS

2
 PLAYER 2

Chicken Game Don’t Swerve

(D)
Swerve
(S)

1 2

Don’t Swerve (D)

1,1 4,2
DD DS
PLAYER 1

3 4
Swerve (S)
2,4 3,3
SD SS

The rows of the matrix represent the available strategies

of player 1.
The columns of the matrix are the strategies of player 2.
Each cell in the matrix indicates one of the 4 possible
states.
2
Chicken Game
1 2
The entries in each cell represent ordinal 1,1 4,2
payoffs in which the first number is the rank DD DS
assigned by player 1 and the second number is
player 2’s ranking.
3 4
A lower number indicates a lower ranking or 2,4 3,3
performance.
SD SS

2
 Definitions for Chicken Game
Set of Players: N={Player 1, Player 2} = {1,2}
Strategy Set for Player 1 (Row Player)
S1 = {Don’t Swerve (D), Swerve (S)} = {D,S}
Strategy Set for Player 2 (Column Player)
S2 = {D,S}
Set of States: S1 × S2 = {DD, DS, SD, SS} = {1, 2, 3, 4}
Payoffs for Player 1
P1={1, 4, 2, 3} (First entry in each cell for states 1 to 4)
Payoffs for Player 2
P2={1,2,4,3} (Second entry in each cell for states 1 to 4)
3
Game Theory Jan 07, 2009 3
 Game in Normal Form

Game Theory Jan 07, 2009 3

Game Theory Jan 07, 2009 3
Game Theory Jan 07, 2009 3
Game Theory Jan 07, 2009 3
Game Theory Jan 07, 2009 3
Game Theory
3
Game Theory Jan 07, 2009 3
Game Theory Jan 07, 2009 3
Game Theory Jan 07, 2009 4
Game Theory Jan 07, 2009 4
 Chicken in Option Form
States
1. Player 1
(1) Swerve N N Y Y
2. Player 2
(2) Swerve N Y N Y
Normal form notation (DD) (DS) (SD) (SS)
State Number 1 2 3 4
Preference Vectors
N Y Y N
Player 1
Y Y N N
2 4 3 1
Y Y N N
Player 2
N Y Y N
3 4 2 1

4
OPTION FORM
Option form was developed to represent more complex
conflict models in a compact and easily understandable
fashion.
Definition: A game in option form is a list of each
player’s options or available courses of action, along with
a rule for specifying the payoffs or preferences.

4
OPTION FORM
An option is a basic action available to a player that can or
cannot be selected.
A Y beside an option means “yes” the option is taken by
the player controlling it.
An N opposite an option means “no” the option is not
selected.
Often 1’s and 0’s are used in place of the Y’s and N’s
respectively.

4
Chicken in graph form

1 2 1 2

3 4 3 4

a) P1 = (1, 4, 2, 3) b) P2 = (1, 2, 4, 3)

4
 THE GRAPH MODEL FOR CONFLICTS

Let N = {1, 2, …, n} denote the set of players and S = {1,

2, …, u} the set of states.
A collection of finite directed graphs Gi = (S, Ai),
i  N, is used to model the conflict where:
(1) The set of vertices S is the set of possible states of the
conflict.
(2) If player i can (unilaterally) move (in one step) from state
k to q, there is an arc with orientation from k to q in Ai.

4
 Exercise 1: Please present the following game in
option form and graph model:

 Normal Form representation – Payoff Matrix

Prisoner 2

Confess Not Confess

Prisoner 1 Confess -3,-3 0,-9
Not Confess -9,0 -1,-1

Game Theory Jan 07, 2009 4

 Example 2: Elmira Conflict
Elmira is located in southwestern Ontario, Canada. In 1989,
the Ontario Ministry of Environment (MoE) detected that the
underground aquifer supplying water for Elmira was
polluted by N-nitroso demethylamine (NDMA). A local
pesticide and rubber manufacturer, Uniroyal Chemical Ltd.
(UR), was identified, since the prime suspect as NDMA is a
by-product of its production line. Hence, a Control Order
was issued by MoE, requiring that UR take expensive
measures to remedy the contamination. UR immediately
appealed to repeal this control order. The Local Government
(LG)} consisting of the Regional Municipality of Waterloo
and the Township of Woolwich, sided with MoE and sought
legal advice from independent consultants on its role to
resolve this conflict. 4
Option Form

DM 1 S7 ≻ S3 ≻ S4 ≻ S8 ≻ S5 ≻ S1 ≻ S2 ≻ S6 ≻ S9

DM 2 S1 ≻ S4 ≻ S8 ≻ S5 ≻ S9 ≻ S3 ≻ S7 ≻ S2 ≻ S6

DM 3 ： S7 ≻ S3 ≻ S5 ≻ S1 ≻ S8 ≻ S6 ≻ S4 ≻ S2 ≻ S9
4
Option Prioritizing for MoE
Preference
Explanation
Statements
MoE most prefers that Uniroyal not abandon its
-4
Elmira plant.
Next, MoE would like Uniroyal to accept the current
3
Control Order.
MoE then prefers that Uniroyal not delay the appeal
-2
process.

-1 MoE would not like to modify the control order.

MoE prefers that Local Government insists that the

5 iff -1 original Control Order be applied (5), if and only if
(iff) it does not modify the Control Order (-1) itself.

50 5
 Graph Model for Elmira Conflict

DM 1 S7 ≻ S3 ≻ S4 ≻ S8 ≻ S5 ≻ S1 ≻ S2 ≻ S6 ≻ S9

DM 2 S1 ≻ S4 ≻ S8 ≻ S5 ≻ S9 ≻ S3 ≻ S7 ≻ S2 ≻ S6

DM 3 ： S7 ≻ S3 ≻ S5 ≻ S1 ≻ S8 ≻ S6 ≻ S4 ≻ S2 ≻ S9
5
Find Nash Equilibrium

5
 Find Nash Equilibrium
 A pair of strategies is a Nash equilibrium if
A’s choice is optimal given B’s choice, and
vice versa.
 Nash equilibrium is a situation, or a
strategy combination of no incentive to
deviate unilaterally.
 A dominant strategy equilibrium is a Nash
equilibrium;

5
Definition of Nash Equilibrium
 A strategy profile is a list (s1, s2, …, sn) of the
strategies each player is using.
 If each strategy is a best response given the other
strategies in the profile, the profile is a Nash
equilibrium.
 Why is this important?
– If we assume players are rational, they will play Nash
strategies.
– Even less-than-rational play will often converge to
Nash in repeated settings.

5
 An Example of a Nash Equilibrium
Column If each strategy
is a best
a b response given
the other
strategies in the
Row a 1,2 0,1
profile, the
profile is a Nash
b 2,1 1,0 equilibrium.
(b,a) is a Nash equilibrium.
To prove this: Given that column is playing a,
row’s best response is b.
Given that row is playing b, column’s best
response is a.
5
 1. Finding Nash Equilibria –
Dominated Strategies
 What to do when it’s not obvious what the
equilibrium is?
 In some cases, we can eliminate dominated
strategies.
– These are strategies that are inferior for every
opponent action.
 In the previous example, row = a is
dominated.

5
An Example of a Nash Equilibrium

Column

a b

a 1,2 0,1
Row
b 2,1 1,0

5
 Example 3: Find the Nash Equilibrium by
eliminating dominated strategies.

Column
a b c

a 73,25 57,42 66,32

Row
b 80,26 35,12 32,54

c 28,27 63,31 54,29

5
  A 3x3 example: Column
a b c

a 73,25 57,42 66,32

Row
b 80,26 35,12 32,54

c 28,27 63,31 54,29

c dominates a for the column player

5
  A 3x3 example:
Column
a b c

a 73,25 57,42 66,32

Row
b 80,26 35,12 32,54

c 28,27 63,31 54,29

b is then dominated by both a and c for the row player.

6
 A 3x3 example:
Column
a b c

a 73,25 57,42 66,32

Row
b 80,26 35,12 32,54

c 28,27 63,31 54,29

Given this, b dominates c for the column player –

the column player will always play b.
6
  A 3x3 example:
Column
a b c

a 73,25 57,42 66,32

Row
b 80,26 35,12 32,54

c 28,27 63,31 54,29

Since column is playing b, row will prefer c.

6
 Column
a b c

a 73,25 57,42 66,32

Row
b 80,26 35,12 32,54

c 28,27 63,31 54,29

We verify that (c,b) is a Nash Equilibrium by observation:

If row plays c, b is the best response for column.
If column plays b, c is the best response by row.
6
Example 4: Find Nash equilibrium for the
following game.

B
Confess Deny
Confess -3*, -3* 0, -5
A Deny -5, 0 -1, -1
6
 Exercise 2: Find Nash equilibrium for the
following game.

Prisoner 2

Confess Not Confess

Prisoner 1 Confess -3,-3 0,-9
Not Confess -9,0 -1,-1

Game Theory Jan 07, 2009 6

 Dominant strategies

 The optimal choice of strategy for one player

no matter what the other player does.

 Method of iterated elimination of

dominated strategies.

6
 Exercise 3: Find the Nash Equilibrium by
eliminating dominated strategies.
 You try this one:
Column
a b c

a 2,2 1,1 4,0

Row
b 1,2 4,1 3,5

6
 Coordination Games
 Consider the following problem:
– A supplier and a buyer need to decide whether
to adopt a new purchasing system.
Buyer
new old

new 20,20 0,0

Supplier
old 0,0 5,5
No dominated strategies!
6
 Buyer
new old
new 20,20 0,0
Supplier
old 0,0 5,5
• This game has two Nash equilibria (new,new) and
(old,old)

No Dominant Strategy for Both

6
Game Theory Jan 07, 2009 7
Game with Pure Strategy
and Mix Strategy

7
 Pure Strategies
• In all previous examples, players are thought of
as choosing to play either one or the other, but
no combination of both; that is, as playing
purely one or the other.
• The strategies presented so far are players’ pure
strategies
• Consequently, equilibria are pure strategy Nash
equilibria.

7
Example 2-4: 1) Determine Nash equilibrium for the
government-beggar game with pure strategy. 2)
Determine Nash equilibrium with mixed strategy.
beggar
Find job Loaf about

Relief 3 ， 2 -1 ， 3
-1 ， 0，0
government
Not 1
relief
No Nash Equilibrium with Pure Strategies.

7
Mixed Strategies
 In many games (such as coordination games) a
player might not have a pure strategy.
 Instead, optimizing payoff might require a
randomized strategy (also called a mixed strategy)

7
 Nash equilibrium with mixed strategies:
x=probability government chooses supporting
y=probability beggar chooses finding a job

Governmnet’s  1  X , Y   3xy  (1) x(1  y )  (1)(1  x) y  0(1  x)(1  y )

payoff ：
 1
 5y 1  0 y  1/ 5
x

 2  X , Y   2xy  3 x (1  y )  (1  x ) y
Beggar’s payoff  2
 (2 x  1)  0
： y
x  1/ 2
X   1 / 2,1 / 2 
Y   1 / 5,4 / 5 
7
Example 2-5: 1) Determine Nash equilibrium for the
husband-wife game with pure strategy. 2) Determine
Nash equilibrium with mixed strategy.
Wife
football shopping

football 2,1 0,0

Husband
shopping 0,0 1,2

(F,F) and (S,S) 　ａｒｅ　 Nash Equilibrium with

Pure Strategies.
7
 Mixed strategy
Instead, each player selects a probability associated with each
action
– Goal: utility of each action is equal
– Players are indifferent to choices at this probability
 a=probability husband chooses football
 b=probability wife chooses shopping they could do better.
Wife
Since payoffs must be equal, for husband:
b*1=(1-b)*2 b=2/3 f s
For wife: Husband f 2,1 0,0
a*1=(1-a)*2 = 2/3 s 0,0 1,2
In each case, expected payoff is 2/3
2/9 of time go to football, 2/9 shopping, 5/9 miscoordinate
If they could synchronize ahead of time they could do better.

7
(F,F) and (S,S) 　ａｒｅ　 Nash Equilibrium with
Pure Strategies.
7
 Supervisory Game

Tax Payer
Tax Evasion Not Tax Evasion a—Money of
Payer
Inspect
Tax Authority c—Cost of Inspect
Not Inspect
F—Fine
Assume c<a+F

7
 Supervisory Game

Tax Payer
Tax Evasion Not Tax Evasion a—Money of
Payer
Inspect a-c+F, -a-F a-c, -a
Tax Authority c—Cost of Inspect
Not Inspect 0, 0 a, -a F—Fine
Assume c<a+F

X-probability for Inspect ， y-probability for Tax avoidness

8
  1
a  c  F a  c  y   ay  Fy  c  0
 1  X , Y   ( x,1  x)     x
 0 a  1  y 
c
y
aF

 2
 a  F  a  y   (a  F ) x  a  0
 2  X , Y   ( x,1  x)     y
 0  a  1  y 
a
x
aF

a
Nash Equilibrium with x 
aF
Mixed strategy ：
c
y 
aF

8
 Ｅｘｃｅｒｓｅｒ : Given game of rock,
paper , and scissors. Find the Nash equilibrium.
Column
rock paper scissors

rock 0,0 -1,1 1,-1

Row
paper 1,-1 0,0 -1,1

scissors -1,1 1,-1 0,0

This game won’t lead to coordination! No Nash equilibrium

with pure strategy.
8
8
Decision Support System for
Zero Sum Game

8
Two-person zerosum games are
a special case of strategic-form
games with n=2.

8
 Two Players, Zero-sum Game

As s ume t hat t wo pl ayer s ar e  and .

St r at egy s et i s S1  {1， 2，
,  m },S2  {1， 2，
,  n }

 a11  a1n 
Payof f mat r i x A＝

   
1 
 ，
 n
a 1
a 
 m1  mn 

i n whi ch aij i s ' s ut i l i t y val ue and ' s val ue i s - aij .

Thi s i s
 m
a zer o- s um game G （S ,S ,A）.
1 2

8
 1

Example 2: Determine the optimal strategy for

two-player’s zero sum game with rational player’s
selection.
1 2 3
1  1 3 2 
 
Gi ven payof f mat r i x A＝ 2  4 3 2 .
 3  6 1 8 

I f I and I I ar e r at i onal , (  2， 3 ) is an equilibrium.

 choos es max min aij   i*
 i j

Rat i onal pl ayer ' s s l ect i on 

 choos es min max aij   j*
j i

8
 1 2 3
 1  1 3  2 
Gi ven payof f mat r i x A＝  2  4 3 2  .
 3  6 1 8 
I f I and I I ar e r at i onal , (  2， 3 ) is an equilibrium.

8
 Optimal Pure Strategy and Saddle

1. If aij*  ai* j*  ai* j , t hen (  i* ,  j* ) i s cal l ed s ol ut i on.

 i* and  j* ar e  and ' s opt i mal pur e s t r at egy，ai* j*
is the value of G，denot ed by VG .

2. Saddl e Poi nt ：I f （ i*， j*）wi t h

ai* j*＝ max min aij＝ min max aij ,
i j j i

t hen（ i*， j*）i s cal l ed s addl e poi nt .

Fr om Exampl e 2, VG  a23，s at i s f yi ng ai 3  a23  a2 j .

8
Exerciser 2: Determine the following zero-sum game’s
saddle point and the optimal strategy’s value.
2
 -2 2 -2 7 VG  3
  3
4 3 8 5
 6 ( i* ,  j* )  ( 2 ,  2 )
 8 -6 2 -1 
8 3 8 7

min max aij  2

 0 2 2  j i

  max min aij  2

 5 4 3  i j

 2 3 4  No s addl e poi nt
 
5 4 2 wi t h pur e s t r at egy.
9
Example 3: Determine the following zero-sum game’s
saddle point and the optimal strategy’s value.
6 5 6 5 5
  -1 VG  5
 1 4 2 -1 
( i* ,  j* )  (1 ,  2 ), (1 ,  4 ),
8 5 7 5 5
  ( 3 ,  2 ), ( 3 ,  4 )
0 2 6 2 0
Saddl e i s not uni que.
8 5 7 5

9
 Sequential Game
 When such a game is sequential it is
sometimes possible to argue that one of the
Nash equilibria is more likely to occur than
the other.
 Using extensive form
 Backward Solution

9
 A Sequential Game Example
Market Dominator
Fight Don’t fight

Stay
out (1,9) (1,8)
Entrant
Enter (0,0) (2,1)
(Enter, don’t fight) and (stay out, fight) are both Nash equilibria
when this game is played simultaneously and we have no way of
deciding which equilibrium is more likely to occur.

9
 A Sequential Game Example
Dominator
Fight Don’t fight
Stay
(1,9) (1,8)
Entrant out
Enter (0,0) (2,1)
Suppose instead that the game is played
sequentially.
We can rewrite the game in its extensive
form.
9
 A Sequential Game Example
Entrant
Stay Enter
out
Dominator Dominator
Fight Don’t Don’t
fight Fight fight
(1,9) (1,8) (0,0) (2,1)
(Stay out, Fight) is a Nash equilibrium. (Enter, Don’t Fight)
is a Nash equilibrium. Which is more likely to occur?

9
9
A Sequential Game Example
 The entrant prefers (Enter, Don’t Fight), but
the market dominator may threat to fight.
 Is the threat credible?
 A market dominator can make it credible.

9
 A Sequential Game Example
Entrant
Stay Enter
out
Dominator Dominator
Fight Don’t Don’t
fight Fight fight
(1,9) (1,8) (0,2) (2,1)
By building up excess capacity, the threat becomes
credible. The potential entrant stays out.

9
Extensive Game

9
Extensive Game

10
 The two-person extensive-form game called
“Market Entry” shown in the following Figure
which contains chance events. The two players
denote two corporations; each one may or may not
product and sell a new product (entry a new
market). Note that DM 1 decides whether to enter
the market first, and that DM 2 learns whether I has
succeeded or failed before DM 2 makes its
decision. Also, DM 2’s production plan is based on
copying DM 1’s, so if DM 1 decides not to entry
the market, DM 2 won’t either. Thus if DM 2 were
to enter the market, it would be in competition with
DM 1. 10
 The two-person extensive-form game called
“Market Entry” shown in the following Figure.

10
Extensive Game

10
Game with Complete Information

10
 Exerciser 2-3: Cournot Duopoly Model

1. If there are two enterprises. Let P=a-(q1+q2) be price,

　 a>0, 　 c － cost ．
enterprise1 ：q1  0, S1  {q1 : q1  0}, 1 (q1 , q2 )  {a  (q1  q2 )  c}q1
enterprise2 ：q2  0, S2  {q2 : q2  0},  2 (q1 , q2 )  {a  (q1  q2 )  c}q2
 1 1
0 q1 * (q 2 )  (a  c  q 2 )
q1 2
 2 1
0 q 2 * (q1 )  ( a  c  q1 )
q 2 2
1 1
q1 *  q 2 *  (a  c)  1 *   2 *  (a  c) 2
3 9

10
2. If there is only an enterprise, then
q  0,  ( q )  ( a  q  c ) q
 1
 0  q*  ( a  c )  q1 *  q 2 *
q 2
1
  *  ( a  c ) 2   1 *  2 *
4
Cournot Duopoly Game is a noncooperateive game that
results in increasing quantities and decreasing profit.

10
 Stackelberg Game

3. If one enterprise enters first (called the leader ),

then  1 ( q1 )  (a  q1  c )q1
 1
 0  q1 *  ( a  c )
q 2
ac
 2 ( q2 )  ( a  q1 -q2  c) q2  ( a  -q2  c ) q2
2

q2*  (a  c) / 4

10
Game with Incomplete Information

10
 Complete Information Model

If there are two enterprises. Let P=A-(q1+q2) be price,

　 a>0, 　 c － cost ．
enterprise1 ：q1  0, S1  {q1 : q1  0}, 1 (q1 , q2 )  { A  (q1  q2 )  c}q1
enterprise2 ：q2  0, S2  {q2 : q2  0},  2 (q1 , q2 )  { A  (q1  q2 )  c}q2
 1 1
0 q1 *( q2 )  ( A  c  q2 )
q1 2
 2 1
0 q2 *(q1 )  ( A  c  q1 )
q 2 2
1 1
q1*  q2 *  ( A  c )  1*   2 *  ( A  c) 2
3 9

10
 Incomplete Information Model

 Given the total quantity ： Q  q1  q2

in which q1 and q2 are quantities for company 1 and company 2,

respectively. P  A  Q  A  (q1  q2 )
 Price function ：

C (q1 )  cq1
 Cost function for company 1 ：

 Cost function for company 2 ：

C (q2 )  cq2

11
 Incomplete Information Model
C (q2 )  cH q2 
 Cost function for company 2 is with probability
Maybe with probability

C (q2 )  cL q2 cH  cL  0 1
 When cost is C (q2 )  cH q2

 2  P  q2  C (q2 )  ( A  (q1  q2 ))q2  cH q2

 When cost is C (q2 )  cL q2

 2  P  q2  C (q2 )  ( A  (q1  q2 ))q2  cL q2

 H A  cH  q1
q2  2

q L  A  cL  q1
 2 2

11
 Incomplete Information Model
C (q2 )  cH q2 
 Cost function for company 2 is with probability
Maybe with probability

C (q2 )  cL q2 cH  cL  0 1

 Company 1 with incomplete information for company 2

 1  P  q1  C (q1 )  ( A  (q1  q2 ))q1  cq1

 1    (( A  q2H  q1 )q1  cq1 )  (1   )  (( A  q2L  q1 )q1  cq1 )

 H A  cH  q1
q2  2

 L A  cL  q1
q2 
 2
 A  c   q2H  (1   )q2L
q1 
 2
11