# L5 - Infinitely Repeated Games We recall that an infinitely...

This preview shows pages 1–9. Sign up to view the full content.

Page 1 of 38 Infinitely Repeated Games We recall that an infinitely repeated game of ,( ),( ) ii G N A is an extensive game with perfect information and simultaneous moves * , , ,( ) i N H P in which      1 { } ( ) t t H A A () P h N for each nonterminal hH * i is a preference relation on A that extends i and satisfies weak separability.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Page 2 of 38 Infinitely Repeated Games We consider three forms of preference relations: Discounting 1 2 3 ( ) ( , , ) t i i i i v v v v is evaluated by 1 1 tt i t δ v Limit of Means 1 2 3 ( ) ( , , ) t i i i i v v v v is evaluated by 1 lim T t i t T vT  Overtaking 1 2 3 ( ) ( , , ) t i i i i v v v v is evaluated by 1 lim T t i t T v 
Page 3 of 38 Nash Equilibrium What are the Nash equilibria of Infinitely Repeated Games?

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Page 4 of 38 Nash Equilibrium Let ,( ),( ) ii G N A . The set of outcomes is i N i AA  . Consider the players play an indefinitely repetition of a cycle of outcomes of length γ , in which each outcome aA is played a β times. a γ β Outcome 1 a is played 1 a β times. Outcome 2 a is played 2 a β times. Outcome A a is played A a β times. Outcome 1 a is played 1 a β times. Outcome 2 a is played 2 a β times. Outcome A a is played A a β times.
Page 5 of 38 Nash Equilibrium If the limit-of-means criterion is used, the payoff of player i in each period, and hence the entire game , is () a β ii γ aA w u a . The payoff profile is hence ( ) ( ) ( ) aa β β i i N i γ γ a A a A iN w w u a u a      . a γ β Outcome 1 a is played 1 a β times. Outcome 2 a is played 2 a β times. Outcome A a is played A a β times. Outcome 1 a is played 1 a β times. Outcome 2 a is played 2 a β times. Outcome A a is played A a β times.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Page 6 of 38 Feasible Payoff Profiles ( ) ( ) ( ) aa β β i i N i γ γ a A a A iN w w u a u a      . A vector () a β γ aA w u a , where a γ β and 0 a β for all a , is called a feasible payoff profile of ,( ),( ) ii G N A u . Note that a feasible payoff profile is not necessarily a payoff profile of ),( ) G N A u .
Page 7 of 38 Nash Equilibrium ( ) ( ) ( ) aa β β i i N i γ γ a A a A iN w w u a u a      But how can we enforce this feasible payoff profile? The feasible payoff profile is enforceable if no player can increase his payoff by deviating! a aA γ β Outcome 1 a is played 1 a β times. Outcome 2 a is played 2 a β times. Outcome A a is played A a β times. Outcome 1 a is played 1 a β times. Outcome 2 a is played 2 a β times. Outcome A a is played A a β times.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Page 8 of 38 Enforceability of Payoffs and Outcomes of Constituent Game Consider the constituent game G .
This is the end of the preview. Sign up to access the rest of the document.

## This note was uploaded on 04/23/2010 for the course CSC CSC5350 taught by Professor Leunghofung during the Winter '09 term at CUHK.

### Page1 / 38

L5 - Infinitely Repeated Games We recall that an infinitely...

This preview shows document pages 1 - 9. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online