空间网络上合作行为演化动力的博弈机制研究
发布时间：2018-08-16 09:00
【摘要】：在自私的群体中合作行为的产生及维持一直是社会科学和自然科学领域内一个重要问题,近些年来吸引了来自心理学,统计物理,行为科学,进化生物学,数学及统计学等众多学科的关注并取得了非常显著的研究成果。尤其是物理学家将物理中的非线性动力学和凝聚态物理的观点引入空间演化博弈及随着复杂网络科学的快速发展更是为这一问题的研究提升到了一个新的高度,另一方面进化生物学家及演化经济学家进行的行为学实验进一步丰富了这一问题的研究。基于上述考虑,本文主要从三个方面对合作行为的演化及维持进行了研究。首先,由于空间网络对合作行为演化的研究取得了非常多的成果,但少有从行为学实验的角度对这一促进效应进行验证的。因此我们从行为学实验的角度设计出空间网络上的重复囚徒困境实验并招募志愿者来参与实验。结果发现,相对于充分混合的情况,空间网络结构确实能够促进合作行为的演化。这是由于处于空间结构中的合作者可以形成团簇,而这些团簇一方面保护了团簇内部的合作者另一方面又抵御了团簇外部背叛者的攻击,因此空间中的合作者可以维持在一定的水平。其次,通过将非对称的适应度计算机制引入空间演化博弈,我们发现相对于传统的空间博弈模型我们的这种模型能够极大的促进合作行为的演化,并且这种促进机制对不同的网络类型都具有普适性。最后,我们研究了空间博弈中的共演化问题。所谓的共演化,就是在博弈过程中,不仅个体的策略更新,同时网络拓扑结构或者个体的一些属性也处于不断的动态变化中。而本文我们基于个体能力的的适应度评估角度对共演化进行了细致的探索,结果发现共演化降低了个体能力的异质性,进而抑制了合作行为的演化。
[Abstract]:The generation and maintenance of cooperative behavior in selfish groups has been an important issue in the social and natural sciences and has attracted interest in recent years from psychology, statistical physics, behavioral science, evolutionary biology, Mathematics and statistics and many other disciplines of concern and achieved very significant research results. In particular, physicists introduce the concepts of nonlinear dynamics and condensed matter physics into the spatial evolutionary game, and with the rapid development of complex network science, the research of this problem has been raised to a new height. On the other hand, behavioral experiments carried out by evolutionary biologists and evolutionary economists further enrich the study of this problem. Based on the above considerations, this paper studies the evolution and maintenance of cooperative behavior from three aspects. First of all, many achievements have been made in the evolution of cooperative behavior in space networks, but few of them are verified from the point of view of behavioral experiments. Therefore, we design the repeated prisoner dilemma experiment on the spatial network from the perspective of behavioral experiment and recruit volunteers to participate in the experiment. The results show that the spatial network structure can promote the evolution of cooperative behavior. This is because the collaborators in the space structure can form clusters, and these clusters protect the cooperators within the clusters on the one hand, and resist the attacks of the external betrayors on the other. Thus the collaborators in space can be maintained at a certain level. Secondly, by introducing asymmetric fitness computer system into spatial evolution game, we find that compared with the traditional spatial game model, our model can greatly promote the evolution of cooperative behavior. And this promotion mechanism is universal to different network types. Finally, we study the co-evolution in space game. The so-called coevolution is that in the process of game, not only the individual strategy is updated, but also the network topology or some attributes of the individual are constantly changing. In this paper, based on the fitness evaluation of individual ability, we explore coevolution carefully. The results show that coevolution reduces the heterogeneity of individual ability, and then inhibits the evolution of cooperative behavior.
【学位授予单位】：云南财经大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O157.5;O225
，
本文编号：2185518
[Abstract]:The generation and maintenance of cooperative behavior in selfish groups has been an important issue in the social and natural sciences and has attracted interest in recent years from psychology, statistical physics, behavioral science, evolutionary biology, Mathematics and statistics and many other disciplines of concern and achieved very significant research results. In particular, physicists introduce the concepts of nonlinear dynamics and condensed matter physics into the spatial evolutionary game, and with the rapid development of complex network science, the research of this problem has been raised to a new height. On the other hand, behavioral experiments carried out by evolutionary biologists and evolutionary economists further enrich the study of this problem. Based on the above considerations, this paper studies the evolution and maintenance of cooperative behavior from three aspects. First of all, many achievements have been made in the evolution of cooperative behavior in space networks, but few of them are verified from the point of view of behavioral experiments. Therefore, we design the repeated prisoner dilemma experiment on the spatial network from the perspective of behavioral experiment and recruit volunteers to participate in the experiment. The results show that the spatial network structure can promote the evolution of cooperative behavior. This is because the collaborators in the space structure can form clusters, and these clusters protect the cooperators within the clusters on the one hand, and resist the attacks of the external betrayors on the other. Thus the collaborators in space can be maintained at a certain level. Secondly, by introducing asymmetric fitness computer system into spatial evolution game, we find that compared with the traditional spatial game model, our model can greatly promote the evolution of cooperative behavior. And this promotion mechanism is universal to different network types. Finally, we study the co-evolution in space game. The so-called coevolution is that in the process of game, not only the individual strategy is updated, but also the network topology or some attributes of the individual are constantly changing. In this paper, based on the fitness evaluation of individual ability, we explore coevolution carefully. The results show that coevolution reduces the heterogeneity of individual ability, and then inhibits the evolution of cooperative behavior.
【学位授予单位】：云南财经大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O157.5;O225
，
本文编号：2185518
本文链接：https://www.wllwen.com/kejilunwen/yysx/2185518.html
