囚徒困境中贴标签控制的促合作设计

Promoting cooperation design of the tag-based donation control in prisoner’s dilemma

中文关键词  囚徒困境  演化稳定策略  机制设计  促合作控制  基于标签捐赠

英文关键词  prisoner’s dilemma  evolutionarily stable strategy  mechanisms-design  promoting cooperation control  tag-based donation

中文摘要

      研究囚徒困境中促进合作演化的控制设计及控制下的合作演化路径问题. 基于标签捐赠机制，设计对个体贴标签的规则，增加个体的认知能力，协助个体决策. 建立反馈控制下的合作演化动态模型, 分析贴标签控制促进合作演化的实现途径, 论证容许控制上界与控制效果之间的一致性. 通过基于合作识别和基于背叛识别两种控制的比较，得到两个结论. 一是合作率较低时, 基于背叛识别的控制更容易促进合作率提升, 反之, 基于合作识别的控制则更容易实现. 二是这两种识别方式需要相同的博弈重复次数. 仿真分析各参数和控制上界对控制律设计和合作演化轨迹的影响.

英文摘要

      The problem of designing a control to promote the evolution of cooperation in the “prisoner’s dilemma” has been investigated and the evolutionary path has been discussed. Design the rules for labeling players as cooperator or defector based on the “tag-based donation” mechanism to improve players’ cognitive ability and help them decision. The replicator dynamics restrained by the feedback control to promoting cooperation has been modeled and the approach of tag-based control to promote the evolution of cooperation has been analyzed. The consistency between the upper bound of the admissible control and the control effect is demonstrated. By comparing the difference between the controls based on recognizing cooperation and based on recognizing defection, the following two results have been obtained. One result is that the control based on recognizing defection is more implementable if the degree of cooperation is low and vice versa. The other result is that the two controls demand same repeated encounter times. The simulations are employed to analyze the effects on the controller and trajectories of the replicator dynamics caused by the parameters of the games and the upper bound of the admissible control.