复杂网络联合生产演化博弈中的历史记忆作用

Influences of Historical Rate of Interest on the Evolutionary Joint Production Game in Complex Networks

研究独立个体在演化博弈中的合作行为一直以来都是个巨大挑战。针对合作演化问题,文中研究了在联合生产演化博弈中,策略更新环节的历史记忆作用对合作涌现的影响。文中引入记忆因子u,对最近两轮回报率进行加权作为策略更新的依据,采用费米动力学通过与邻居配对比较进行策略更新。利用蒙特卡罗仿真,先后分析了记忆作用和固定成本下合作频率与利润系数的关系图、合作频率随时间演化的曲线图以及记忆作用下网络机制与合作频率的关系图。研究发现,合作频率的提升与记忆权重取值正相关,增加历史记忆作用能够显著促进群体合作并降低出现稳定比例合作者所需的利润系数阈值,且在BA网络、ER随机网络和二维格子网络上具有普适性。

Researching on the cooperative behavior of independent individuals in evolutionary games has always been a challenge.In view of the problem of cooperative evolution,the effect of memory in strategy updating on the emergence of cooperation in joint production game is studied in this paper.The memory factor u is introduced,and the weighted average return of interest in the nearest two rounds is adopted as a basis for strategy update.Then,Fermi dynamics is used to update strategies by pairwise-comparison.Monte Carlo simulation is applied to analyze the relationship between cooperative rate and profit coefficient under the effect of memory factor and fixed cost,and the evolution curve of cooperative rate over time.The relationship between network mechanism and cooperative rate under the memory factor are analyzed.The proposed study finds that the increase in cooperative rate is positively related to the value of memory weight.Additionally,increasing the effect of historical memory can significantly promote group cooperation and reduce the threshold of profit coefficient required for stable fraction of cooperators to appear.This law is also universal on BA network,ER stochastic network and two-dimensional lattice network.