Research On Conditional Neutral Punishment And Mixed Evolution Effective Of Network Game

How to understand the emergence and maintenance of group cooperation under natural selection is a huge challenge in artificial intelligence and social sciences.Evolutionary game theory was proposed and have been effective theoretical tools for studying group cooperation behavior.In nature and human society,the interaction between individuals constitutes a complex system.To describe this complex system and characterize the cooperative behavior between groups more conveniently,complex networks are introduced into the evolutionary game.In this paper,we propose two mechanisms to promote group cooperation based on evolutionary game theory on complex networks.and the research work is as follows:(1)Based on punishment mechanism plays an important role in promoting cooperation.In real society,individuals tend to punish other players based on certain conditions rather than punish them directly.Thus,we introduce a conditional punishment mechanism and study how this mechanism affects the evolution of cooperation.Namely,an individual can punish his/her neighbors with the opposite strategy when his/her payoff is lower than the average payoff of his/her neighbors.The simulation results show that this mechanism promotes cooperation effectively even with antisocial punishment.By adopting such a mechanism,cooperative punishers form shields to protect cooperators inside,while defective punishers hide behind defectors without punishing anyone.(2)Based on the network reciprocity mechanism,that is,all agents participating in the game can adaptively adjust the strength of the relationship with other agents,group cooperation behavior can be promoted.In this paper,by introducing some agents with adaptive adjustment ability,the relationship between the proportion of adaptive adjustment individuals and the emergence of group cooperation behavior is studied.Experiments have found an optimal ratio of adaptive agents,under which cooperative behavior is most promoted.The primary reason is that in the process of evolution,homogeneous individuals gradually evolve into heterogeneous populations.However,the heterogeneity of the population has led to the emergence of multiple strategy cycles(similar to that in the rock-scissors-cloth).This result can provide a theoretical basis for the optimal deployment ratio of agents in the multi-agent collaboration problem.