Cooperation levels of iterated Prisoner's dilemma game with punishing agents and strategy update processes

Abstract

Modern societies exhibit high cooperation dominance, as opposed to defection with respect to observing rules set by the society. A factor that may have contributed to this behavior is the presence a law-enforcing body implementing taxes for citizens and fines for those who defect. This study introduces the addition of a third class of agent: the punishers used to model law-enforcement. Cooperators and defectors in iterated lattice-based Prisoner's dilemma games are allowed change strategy based on global conditions or local interactions. We determine the punishing agent density from single-shot games, leveling the pay-offs of cooperators and defectors at 0.13. Using this density, we simulate iterated games for agents on the lattice followed by strategy updates. Results from this study show a 123% density increase for lattices with 0.13 as opposed to lattices with no punishing agents, with cooperators and defectors allowed to change strategy based on global conditions. Local interactions on the other hand does not promote cooperation in lattices with no punishing agents, while the addition of 0.13 punishing agents results to a 181% cooperator density increase between these conditions. Over-all, we show how the addition of punishing agents and allowing for strategy changes promotes higher levels cooperation.