引用本文:王龙,吴特,张艳玲.共演化博弈中的反馈机制[J].控制理论与应用,2014,31(7):823~836.[点击复制]
WANG Long,WU Te,ZHANG Yan-ling.Feedback mechanism in coevolutionary games[J].Control Theory and Technology,2014,31(7):823~836.[点击复制]
共演化博弈中的反馈机制
Feedback mechanism in coevolutionary games
摘要点击 4347  全文点击 2688  投稿时间:2014-01-22  修订日期:2014-05-15
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/CTA.2014.40059
  2014,31(7):823-836
中文关键词  反馈机制  演化动力学  溯祖理论  机制设计  通牒博弈
英文关键词  feedback mechanism  evolutionary dynamics  coalescent theory  mechanism design  ultimatum game
基金项目  国家自然科学基金资助项目(61020106005, 61375120).
作者单位E-mail
王龙* 北京大学 系统与控制研究中心 longwang@pku.edu.cn 
吴特 西安电子科技大学 机电工程学院  
张艳玲 北京大学 系统与控制研究中心  
中文摘要
      本文研究共演化动力学在合作演化中的作用. 系统的状态由个体的策略决定. 系统状态与个体的属性共同演化, 形成一种反馈机制. 特别是当个体能够根据博弈的结果调整社会关系时, 这种调整势必影响未来的博弈. 这种反馈机制在适当的时间尺度下, 总是能够促进合作. 首先, 分析了个体策略与群组属性共演化的情形, 在基于个体选择的层面上, 结合溯祖理论和演化集合论, 给出了区域性利他行为涌现的条件. 其次, 给出了结构群体中策略选择的参数判据, 将两策略判据和适应动力学结合起来, 发现通过调整与收益矩阵无关的参数可以实现性能控制. 最后, 研究了不同的角色分配方案在最后通牒博弈中对公平行为演化的影响, 发现当个体的先行者优势或所配置的资源依赖于先前分配结果时, 公平的分配方案及对公平的要求得以建立. 共演化这种反馈机制在用博弈论解决编队控制、资源配置方案设计等实际问题时显示出广泛的应用前景.
英文摘要
      We investigate the effects of coevolutionary dynamics on the evolution of cooperation. The state of the system is described by the collective level of individual’s strategies. The state coevolves with individual’s property, constituting a feedback mechanism. In particular, whenever individuals are allowed to adjust their social ties in accordance with the game outcome, this adjustment is predestined to affect the payoff in the future game interactions. This feedback mechanism proves effective in promoting cooperation under appropriate time scale of strategy updating to game happening. Next, there comes the coevolutionary dynamics of individual’s group affiliation and strategies. On the basis of individual-level selection, by the combination of coalescent theory and evolutionary set theory, the conditions are derived for the parochial altruism to establish. We present the criterion of strategy selection in the structured populations for multiple strategies. We apply this criterion for two-strategy setting in the framework of adaptive dynamics, and find that by adjusting the game matrix independent criterion parameters, certain performance of control systems can be realized. In the ultimatum game, whenever the first mover assignment or the size of the resource to be divided is dependent on the outcome of previous allocation rule, the fair allocation rule and the egalitarian asking demand can be stabilized. The coevolutionary dynamics, as a feedback mechanism, promises prosperous applicability in resolving the problems such as the multi-agent formation and the mechanism design in resource allocation.