基于期望的概率休眠促进合作的演化

Evolution of Expectation-Based Probabilistic Dormancy Promote Cooperation

摘要: 针对演化博弈中合作的产生与演化，本文提出了一种基于节点期望的概率休眠机制演化模型。我们重点研究了期望与休眠机制对节点策略行为和群体合作的影响，在方格网络上仿真验证了该模型。仿真结果表明，只有节点的期望收益在一个范围时系统中合作现象才会出现，并且存在一个最佳期望收益可以使系统的合作率最高。节点休眠时间太短和太长不利于合作的演化，不同的背叛诱惑存在最佳的休眠时间可以使得合作率最高。存在一个最优的参与博弈参数可以使得系统合作率最高，合作现象无法在小的参与博弈参数时存在，在参与博弈参数很大时即使背叛诱惑很大，系统中依然存在一个较低的合作率。通过仿真验证，我们发现基于期望的概率休眠机制可以有效的促进合作的产生与演化。

Abstract: In response to the generation and evolution of cooperation in evolutionary games, this paper pro-poses a probabilistic dormant mechanism evolutionary model based on node expectations. We fo-cus on the effects of expectation and dormancy mechanism on node strategy behavior and group cooperation, and validate the model by simulation on square lattice network. The simulation results show that cooperation occurs in the system only when the expected payoffs of the nodes are in a range, and that there exists an optimal expected payoff that can lead to the highest cooperation rate in the system. The node dormancy time is too short and too long for the evolution of cooperation, and the existence of an optimal dormancy time for different betrayal temptations can make the co-operation rate highest. There exists an optimal participation game parameter that can make the system have the highest cooperation rate, the cooperation phenomenon can not exist at small par-ticipation game parameter, and even if the betrayal temptation is large when the participation game parameter is large, there still exists a low cooperation rate in the system. Through simulation verification, we find that the expectation-based probabilistic dormancy mechanism can effectively promote the generation and evolution of cooperation.