Memory-based evolutionary prisoner’s dilemma game with neighbor screening mechanism

Memory effects of individuals have been demonstrated to significantly promote cooperation, attracting widespread attention among scholars exploring the underlying dynamics of cooperative behavior. In this paper, we build a strategy updating framework by proposing a neighbor screening mechanism and combining particle swarm optimization algorithm in memory-based evolutionary prisoners’ dilemma game. Under the proposed mechanism, individuals may not imitate a neighbor with the only highest payoff from the current round when updating their strategy. Instead, individual evaluates the historical performance by his/her neighbors’ the frequency of choosing cooperation strategy and being chosen as an optimal neighbor and screen out the neighbors who do not meet the defined initial threshold. Moreover, a threshold adjustment parameter α is introduced to strengthen the flexible of the threshold. In addition, the proposed mechanism is compared in two situations: fixed memory and dynamic memory. Specifically, in the neighbor screening mechanism with dynamic memory, each individual’s memory will decay or remain unchanged according to the relationship between their current payoff and average payoff of all players in the local group. The simulation results reveal that the advantages of short memory length and dynamic memory in promoting cooperation. Furthermore, the synergistic effect between initial threshold and memory length better promotes cooperation. Additionally, a slight increase in the threshold adjustment parameter α promotes cooperation when the initial threshold is low. These findings shed light on how cooperation can be enhanced through specific rules.