Cooperative behavior under the influence of multiple experienced guiders in Prisoner's dilemma game

In evolutionary game theory, the emergence and maintenance of group cooperative behavior is usually challenged by the lure of high-payoff defection behavior.Recently, besides imitation rules, the adaptive ability of individuals under limited information is also the key to adjust strategies, for example, individuals based on reinforcement learning rules by judging whether previous performance is satisfactory. In realistic scenarios, individuals with rich experience usually lead those who are inexperienced and play a guiding role in the group. Here we propose a multi-guider game model. Players on each layer of the network play different roles and follow different strategy update rules. Specifically, guiders use reinforcement learning rules to update their strategies in the upper network, and guided players use payoff-based imitation rules to update their strategies in the lower network. As the evolution progresses, guided players in the lower layer begin to reference the experienced guiders in the upper layer to update their strategies. A large number of Monte Carlo simulation results show that inexperienced individuals in the group are able to learn from the experience of others with the experienced guidance of multiple guiders. In addition to the improvement of group decision-making, the cooperative behavior can also be maintained at a higher level in the simulation of social dilemma.