The Competitions of Forgiving Strategies in the Iterated Prisoner's Dilemma

Abstract

The iterated prisoner's dilemma or IPD game has been widely used in modelling interactions among autonomous agents. According to the tournament competitions organized by Axelrod, Tit-for-Tat emerged as the most effective strategy on the assumption of an environment clinically free of communicative error or noiseless. However, with noise present, Tit-for- Tat contradictorily finds itself more difficult to maintain cooperation. In this study, the competitions of our proposed strategies and other Tit-for- Tat like strategies in the environment with different levels of noise are presented. The main result is that our proposed strategies provide the most effective performance in both round-robin tournaments and evolutionary dynamics.