Upload
others
View
12
Download
0
Embed Size (px)
Citation preview
Introduction to Game Theory
Introduction to Game Theory
March 7, 2018
Introduction to Game Theory
Introduction
Basic Concepts in Noncooperative Game Theory
Actions (welfare or profits)
Help us to analyze industries with few firms
What are the firms’actions?
Two types of games:
1 Normal Form Game2 Extensive Form game
Two types of actions: Pure and Mixed
Information: Perfect and Imperfect
Introduction to Game Theory
Introduction
Examples of Noncooperative Game Theory
Introduction to Game Theory
Introduction
Normal Form Games
Definition 2.1: A normal form game:
1 N players whose names are listed in the set I ≡ {1, 2, ...,N}2 Each player i , i ∈ I , has an action set Ai , whereAi = {ai1, ai2, ..., aiki }
3 List of actions chosen by each player:a ≡ (a1, a2, ..., ai , ..., aN )
4 Each player has a payoff function πi ∈ R
Introduction to Game Theory
Introduction
Normal Form Game
"Peace-War Game" (Prisoners’Dilemma)
Country 1
Country 2War Peace
War 1, 1 3, 0Peace 0, 3 2, 2
Let us apply Definition 2.1.........
Introduction to Game Theory
Introduction
Equilibrium Concepts
We would like to obtain one outcome (unique eq.)
Outcome of the game: a ≡ (a1, a2, ..., ai , ..., aN )a−i ≡ (a1, ..., a−i , ai+1, ..., aN )
let’s talk about a−i
Hence, an outcome a can be expresses as a ≡ (ai , a−i )
Introduction to Game Theory
Introduction
Equilibrium in dominant actions
Definition: A particular action ai ∈ Ai is said to be adominant action for player i if no matter what all otherplayers are playing ai always maximizes i’s payoff
πi (ai , a−i )
for every ai ∈ Ai .Example: War-Peace gameWhat is a Dominant Strategy for player 1?
Country 1
Country 2War Peace
War 1, 1 3, 0Peace 0, 3 2, 2
Note that an outcome is always composed by a DominantStrategy
Introduction to Game Theory
Introduction
Payoff matrix (Normal Form Game)
Firm A
Firm BLow Prices High Prices
Low Prices 5, 5 9, 1High Prices 1, 9 7, 7
Low prices yield a higher payoff than high prices both when afirm’s rival chooses low prices and when it selects high prices
Low prices is strictly dominant strategy for both firmsHigh prices is referred to as a strictly dominated strategy
Introduction to Game Theory
Introduction
A strictly dominated strategy can be deleted from the set ofstrategies a rational player would use.
This helps to reduce the number of strategies to consider asoptimal for each player.
In the above payoff matrix, both firms will select “low prices”in the unique equilibrium of the game.
However, games do not always have a strictly dominatedstrategy.
Introduction to Game Theory
Introduction
Battle of the Sexes (coordination games)
Jacob
RachelOpera Football
Opera 2, 1 0, 0Football 0, 0 1, 2
Introduction to Game Theory
Introduction
Battle of the Sexes (coordination games)
No Dominant Strategies!
Hence, there does not exist an equilibrium in dominant actions
Introduction to Game Theory
Introduction
Nash Equilibrium (NE)
Definition: An outcome a = (a1, a2, ..., ai , ..., aN ) is said to bea NE if no player would find it beneficial to deviate providedthat all other players do not deviate from their strategiesplayed at the Nash outcome
πi (ai , a−i ) ≥ πi (ai , a−i )
for every ai ∈ Ai .An equilibrium in dominant action is a NE but a NE ; eq.D.A
Introduction to Game Theory
Introduction
Nonexistence of a Nash Equilibrium
After 30 years of marriage................. ;)
Jacob
RachelOpera Football
Opera 2¯, 0 0,2
¯Football 0,1¯
1¯, 0
Introduction to Game Theory
Introduction
Best-Response functions to solve for NE
Definition: in a two-player game, the BRF of player i is thefunction R i (aj ), that for every given action aj of player jassigns an action ai = R i (aj ) that maximizes player i’s payoffπi (ai , aj )
Example: Battle of the sexes
Introduction to Game Theory
Introduction
Battle of the Sexes (coordination games)
Jacob
RachelOpera Football
Opera 2, 1 0, 0Football 0, 0 1, 2
Example Battle of the sexes
RJ (aR ) ={
Opera if aR = OperaFootball if aR = Football
RR (aJ ) ={
Opera if aJ = OperaFootball if aJ = Football
Then if a is a NE , then ai = R i (a−i ) for every player i .