Evolutionary game theory in the cognitive sciences

Game theory models something very relevant to psychologists (in particular social psychologists): conflict and cooperation between decision-makers. Unfortunately, classical game theory demands that these decision makers are rational (in a mathematically precise sense). This definition of rationality is challenged empirically (by work like Tversky & Shafir) and on a theoretical level (by complexity results on finding or approximating Nash equilibria: a PPAD-complete problem).

Economists and mathematicians (and others) have taken two approaches to overcome this problem. The first is top-down approach of limiting the agent's abilities from an all-powerful rational agent down; this is the bounded rationality approach. The other is the bottom-up approach of evolutionary game theory: start with the simplest agents (that can't even make decisions) and have natural selection, imitation, or another simple dynamic process evolve the population over time. This seems to dove-tail nicely with the biogenic approach to cognition.

Are there examples of the tools of evolutionary game theory being used in social psychology or other sub-disciplines of the cognitive sciences? Is there a survey (or book) on EGT's impact on the cognitive sciences?

Game theory strategies

John von Neumann’s minimax theory

Posted on June 7, 2012 by Barry Hughes

“Keeping up with him was all but impossible. The feeling was you were on a tricycle chasing a racing car”

Israel Halperin on John von Neumann

John von Neumann made huge contributions to many areas of mathematics and was also one of the founders of game theory.

His mathematical ability shone through from a young age. At the age of six he was able to divide two eight-digit numbers in his head and by the age of eight he had mastered calculus.

One of his major contributions to game theory was the Minimax theorem, which he proved in 1928.

 

The basic idea applies to two player zero-sum games. As it is a zero-sum game one player’s payoff is the opposite of the other’s payoff. So if one gets a payoff of +10 then the other will get a payoff of -10, the two payoffs add up to zero.

In any game you are trying to maximise your own payoff and your opponent is trying to do the same thing. In a zero-sum game, if you maximise your payoff then that is the same as minimising your opponent’s payoff. There is only a fixed amount to be shared between the players in a zero-sum game so if one wins an amount then the other loses an equal amount.

Given that you know that your opponent is going to play a strategy that will maximise their payoff, you want to play the strategy against them that will minimise their payoff (and therefore maximise yours).

Putting these two things together means that you want to play the strategy that minimises your opponent’ s maximum payoff. This is where the name ‘minimax’ comes from.

In zero-sum games the minimax solution is the same as the Nash equilibrium.