惩罚因子在重复囚徒困境博弈中的研究

RESEARCH ON PENALTY GENE IN ITERATED PRISONER'S DILEMMA GAME

针对囚徒困境博弈中收益矩阵参数无法动态更新的问题,提出一个带惩罚因子的囚徒困境博弈模型。该模型中的个体可以根据自身的策略,动态修改收益矩阵中的参数,在每轮博弈之后个体根据邻居收益更新自己的策略。仿真结果表明,惩罚因子可以有效地促进合作策略的涌现,另外发现,合作策略的涌现不仅与惩罚因子有关,而且与网络个体的初始策略有关。当社团内部的个体采取相同策略,社团之间采取不同策略时,更有利于合作策略的涌现。最后,惩罚因子还可以提高参与者的平均收益。

Aiming at the problem that the parameters of payoff matrix in prisoner＇ s dilemma game cannot update dynamically, we put forward a prisoner＇ s dilemma game model with penalty gene, in which the individuals can dynamically modify the parameters of payoff matrix according to its own tactic. The individuals can also update its tactic according to its neighbours＇ payoff after one round of game. Simulation re- sults show that the penalty gene can efficiently boost the emergency of cooperation tactic. On the other hand, we also find that the emergency of the cooperation tactic is not related to the penalty gene only but also to initial tactics of the individuals of network. When the individuals in the community adopt the same tactics but different communities adopt different tactics, the cooperation tactics are more tend to come forth. At last, the penalty gene can also promote the average payoffs of players.