Cooperation of a Dissatisfied Adaptive Prisoner's Dilemma in Spatial Structures

We study the cooperative behavior of a dissatisfied adaptive prisoner's dilemma via a pair updating rule. We compare two kinds of relationship among the competing agents, one is the well-mixed population and the other is the two-dimensional square lattice. It is found that the cooperation emerges in both the cases and the frequency of cooperation is enhanced in the square lattice. Though it is impossible for the cooperators to have a higher average payoff than that of the defectors in the well-mixed case, the cooperators in the spatial square lattice could have higher average payoffs in certain regions of the game parameters. We theoretically analyze the well-mixed case exactly and the square lattice by pair approximation. The theoretic results are in agreement with the simulation data.