Which Cooperative Strategies are Evolutionarily Stable?

Abstract

In repeated games, reciprocity helps to sustain group cooperation. Using the Markovian strategy as a touchstone, we explore the evolution of cooperation in the infinitely repeated and noisy n-person Prisoner's Dilemma. With some modified game settings, we show that a class of cooperative strategies based on reciprocity may be evolutionarily stable. When a cooperative strategy satisfies some properties in our study, that strategy becomes evolutionarily stable if it outperforms the mutant permanent defector. We also argue that evolutionary stability in a noisy n-person setting implies the Markov perfect equilibrium.