Abstract:
In the past two decades the entropy method has been successfully employed in the study of repeated games. I will present a few results that demonstrate the relations between entropy and memory. More specifically: a finite game is repeated (finitely or infinitely) many times. Each player $i$ is restricted to strategies that can recall only the last $k_i$ stages of history. The goal is to characterize the (asymptotic) set of equilibrium payoffs. Such a characterization is available for two-player games, but not for three players or more.
Related papers: