Iterated prisoner's dilemma with choice and refusal of partners: Evolutionary results

Abstract

In a series of papers we have examined what happens when individuals make very calculated choices of partners, based on past interaction histories [17, 1, 16]. In Iterated Prisoner's Dilemma with Choice and Refusal (IPD/CR), players use expected payoffs, which are based on the play history between the players plus an initial expectation, to assess the relative desirability of potential partners and refuse play with those judged to be intolerable. We have primarily studied this model using evolved populations of finite state machines. In each generation, individual behaviors generate social networks of interacting players. Here we provide an overview of our previous evolutionary results, and include some preliminary results on the impact of increasing the population size and including more randomness into the partner selection procedure.

Special thanks to Leigh Tesfatsion, who helped develop the IPD/CR model and with whom we have had many thoughtful discussions. Partial support came from an Iowa State Research Grant funded under DHHS Grant # 2SO7RR07034-26 and from the Los Alamos Center for Nonlinear Studies. Many thanks to the Iowa State Physics and Astronomy Department, especially the Gamma Ray Astronomy research group, for providing computer facilities and office space to Mark Smucker.