On prisoner’s dilemma game with psychological bias and memory learning

Abstract

Since we live and cooperate in a variable and complex network of relationships, the intricate interactions between individuals emerge amazing complex population dynamics. Combining self-based memory learning, neighbors-based learning and ϵ−greedy strategy selection, we here propose a prisoner’s dilemma game model with diversified behavior change. Results show that individuals’ cooperation behavior could be restrained while the underlying contact network structure with possessing a high average connection. In the case of a small value of psychological bias factor, the ϵ−greedy strategy selection way could just be conducive to better stimulate the individuals’ cooperation behavior in a homogeneity network. However, in the case of a large value of psychological bias factor, ϵ−greedy strategy selection way can enhance the propensity of cooperation regardless of the contact network structure. More comparative analysis of individuals’ gain in a heterogeneity network and a homogeneity network, we find that one’s average gain in former network is significantly greater than that in the latter network. In addition, an optimal value of psychological bias factor exists which can realize benefits maximization in the population while individuals tend to choose strategies at random. Finally, we show that no matter what strategy a player adopted, his average gain shows a significant linear relation with his connection. We try to provide some schemes that can effectively promote the mutual cooperation between people, and offer some reasonable research perspective for the study of evolutionary game related to individuals’ rational.