基于特殊个体的网络进化博弈群智合作激发调控  

Excitation and control of group intelligence cooperation in network evolutionary game based on special individuals

作　　者：王浩淼 魏晨 邓亦敏 WANG HaoMiao;WEI Chen;DENG YiMin(Bio-inspired Autonomous Flight Systems(BAFS)Research Group,Science and Technology on Aircrafi Control Laboratory,School of Automation Science and Electrical Engineering,Beihang University,Beijing 100083,China)

机构地区：[1]北京航空航天大学自动化科学与电气工程学院,飞行器控制一体化技术重点实验室仿生自主飞行系统研究组,北京100083

出　　处：《中国科学：技术科学》2023年第2期167-176,共10页Scientia Sinica(Technologica)

基　　金：科技创新2030-“新一代人工智能”重大项目(批准号:2018AAA0102403);国家自然科学基金项目(批准号:U20B2071,91948204,U1913602,U19B2033)资助。

摘　　要：引导群体中智能行为的产生与传播是群体智能领域的重要研究方向.相互合作是自然界中常见的群体智能行为.本文建立了一种基于特殊个体的网络进化博弈模型,分析了群体激发合作水平的进化特性以及调控手段.首先通过在规则网络进化博弈模型中建立有效的干预控制体系结构,实现对群集系统的调控作用;然后通过设计基于群体智能的特殊个体控制规则,特殊个体能起到促进合作行为涌现的群智激发作用;最后利用蒙特卡罗仿真方法分析了特殊个体对群体进化特性的影响,并讨论了不同参数条件对合作水平的影响.Guiding the generation and propagation of intelligent behavior in crowds is an important research direction in the field of swarm intelligence. Cooperation is a common behavior of swarm intelligence in nature. In this study, an evolutionary game model based on special individuals is established to analyze the evolutionary characteristics and regulatory measures of group-inspired cooperation levels. First, an effective intervention control architecture is established in the rule network evolutionary game model to realize the regulation effect on the cluster system. Second, by designing the control rules of special individuals based on swarm intelligence,special individuals can stimulate the emergence of cooperative behaviors. Finally, the Monte Carlo simulation method is used to analyze the influence of special individuals on the evolution characteristics of a group, and the influence of different parameter conditions on the cooperation level is discussed.

关 键 词：特殊个体 网络进化博弈 软控制 群智合作激发 蒙特卡罗仿真 

分 类 号：O225[理学—运筹学与控制论]