Could feedback-based self-learning help solve networked Prisoner´s Dilemma?

We present a self-learning evolutionary Prisoner´s Dilemma game model to study the evolution of cooperation in network-structured populations. During the evolutionary process, each agent updates its current strategy with a probability depending on the difference feedback between its actual score and score aspiration. Each agent´s score is a weighed mean of its payoff coming from its neighbors (social partners) and the payoff of its social partners obtaining from it. Simulation results show that the cooperation level in the structured populations increases with increasing the weight of partners´ obtaining payoff in the score. More interestingly, we find that very similar evolution of cooperation can respectively emerge in lattice, small-world and scale-free networks under the learningfeedback updating rule. Moreover, we provide theoretical analysis and qualitative explanations for these numerical simulations. Our work may provide an effective way to solve the dilemma of cooperation for structured populations.