Abstract: With time going by，two nodes in complex network may adopt some actions for some conflicts of interests and they may either cooperate or detect their neighbors.For the node who detects most，its neighbor tends to cut their relationship and reconnect a new neighbor，this situation fit for real world.This paper puts forward a new pseudo value first algorithm based on tit-for-tat strategy in game theory and studies the network statistic characters with the network nodes unchanged.The simulation results show that the new algorithm can optimize the community structure，does not change scale free character of network and degree related coefficient is lower than zero that is to say the node who has large degree declines to connect the node who has small degree.

Key words: prisoner’s dilemma, tit-for-tat, Nash equilibrium, complex network

摘要： 复杂网络中的两个节点，随着时间的推移，由于利益冲突，可能会采取一些行动，合作或者背叛，对于背叛过多的节点，需要断开重连，这也恰好反映了现实情况。根据重复博弈中的tit-for-tat策略，提出了一种伪度优先算法，研究了在不改变节点个数情况下，复杂网络的统计特性。仿真结果表明，该算法可以优化网络的社团结构，不改变网络的无标度特征，经过演化的网络度相关系数小于0，说明度大的节点倾向于与度小的节点连接。

关键词: 囚徒困境, tit-for-tat, Nash均衡, 复杂网络