重复性公共品博弈中的合作演化

发布时间：2018-02-01 14:24

本文关键词： 混合均匀群组公共品博弈惩罚与奖励行为多样性针锋相对非线性惩罚阈值　出处：《西安电子科技大学》2015年硕士论文　论文类型：学位论文

【摘要】：生物之间的竞争无处不在,达尔文的自然选择理论表明在竞争中适应度大的个体更有可能存活下来。不存在任何机制的情况下,合作个体的适应度低于不合作的个体。从而,为了生存,个体在竞争中倾向于选择不合作。然而,生物系统最基本的一个特征就是合作。对于很多不同层次的生物组织,从单细胞生物到动物群组,合作的交互行为都是必需的。因此,竞争性的世界中合作行为的广泛存在就令人疑惑。从而,如何解释自私个体间合作行为的涌现以及长期以来合作行为的维持受到了诸多领域科学家的关注。演化博弈论为上述问题的研究提供了有利的理论依据。本文致力于探讨基于混合均匀群组的重复性公共品博弈的合作演化问题,主要成果如下:首先,在混合均匀种群中引入惩罚和奖励机制后,研究群组中不同策略的演化。在博弈之前每个个体需要付出一定代价来表明自己对于合作者或背叛者的态度,惩罚.或者奖励,而不考虑之后自己的策略。惩罚和奖励单独作用均不能够促进合作的演化,而在惩罚和奖励机制的共同作用下,合作水平随选择强度的增加而提高。研究结果还表明,在惩罚和奖励共存的博弈中,奖励比惩罚会有更大的优势。其次,在考虑噪声的情况下,根据行为的多样性提出了一个基于混合均匀群组的新模型来探究自适应策略在重复性公共品博弈中的演化。我们引入针锋相对特性,合作概率代表行为的多样性,用基于频率选择的莫兰过程来刻画群组的动态演化过程,这一过程表明了大自然的选择:越是成功的个体越容易被复制。研究表明,在考虑行为多样性的情况下,即使存在噪声的影响,针锋相对机制依然能够促进合作的演化。再次,在上述模型的基础上我们考虑非线性和不同的群组规模对合作演化的影响。我们知道不同的非线性类型对重复性公共品博弈中合作演化的影响不尽相同。结果表明,在行为多样性和针锋相对特性存在的情况下,协同促进作用的非线性会促进合作的演化,而折扣效应的非线性会抑制合作的演化。但是,在折扣效应的非线性公共品博弈中,存在最适宜合作演化的群组规模值。最后,构建了一个相对简单的理论模型来研究当惩罚行为基于某个特定阈值时,各个策略在混合均匀群组的公共品博弈中是如何演化的。在每一轮的博弈中,个体付出不同的代价,而惩罚只是针对群组中投入低于某一特定值的个体。研究表明,当惩罚阈值适中时,参加公共品博弈的个体的总投入量会处于最大值,并且出现频率最高的策略类型是投入量略高于惩罚阈值的策略。
[Abstract]:Competition between organisms is ubiquitous, and Darwin's theory of natural selection suggests that individuals with greater fitness in competition are more likely to survive without any mechanism. The fitness of cooperative individuals is lower than that of non-cooperative individuals. Therefore, in order to survive, individuals tend to choose non-cooperation in the competition. One of the most basic features of biological systems is cooperation. For many different levels of biological tissue, cooperative interaction is required from single-celled organisms to animal groups. The widespread existence of cooperative behavior in a competitive world is puzzling. How to explain the emergence of cooperative behavior among selfish individuals and the long-term maintenance of cooperative behavior have attracted the attention of scientists in many fields. Evolutionary game theory provides a favorable theoretical basis for the study of these problems. The cooperative evolution of repeated public goods game based on mixed uniform group is discussed. The main results are as follows: first, after the introduction of punishment and reward mechanism in the mixed uniform population. Study the evolution of different strategies in groups. Each individual has to pay a price before the game to show his or her attitude towards collaborators or collaborators, punishment, or rewards. Punishment and reward alone can not promote the evolution of cooperation, but under the combined action of punishment and reward mechanism. The level of cooperation increases with the increase of the intensity of selection. The results also show that in the game of coexistence of punishment and reward, reward will have more advantages than punishment. Secondly, considering noise. According to the diversity of behavior, a new model based on mixed uniform group is proposed to explore the evolution of adaptive strategy in repetitive public goods game. We introduce the tit-for-tat characteristic. The cooperative probability represents the diversity of behavior. The Moran process based on frequency selection is used to describe the dynamic evolution of the group. This process shows nature's choice: the more successful individuals are, the easier they are to be replicated. Studies have shown that, even when noise is involved, the diversity of behavior is considered. Tit-for-tat mechanism can still promote the evolution of cooperation. On the basis of the above model, we consider the influence of nonlinearity and different group size on cooperative evolution. We know that different nonlinear types have different effects on cooperative evolution in repeated public goods games. Ming. In the presence of behavioral diversity and tit-for-tat characteristics, the nonlinearity of cooperative facilitation will promote the evolution of cooperation, while the nonlinearity of discount effect will inhibit the evolution of cooperation. In the nonlinear public goods game with discount effect, there is a group size value that is most suitable for cooperative evolution. Finally, a relatively simple theoretical model is constructed to study when the punishment behavior is based on a specific threshold. How each strategy evolves in the public goods game of mixed uniform groups. In each round of games, individuals pay different costs. However, the punishment is only for the individuals in the group who invest less than a certain value. The research shows that when the penalty threshold is moderate, the total input amount of the individuals participating in the public goods game will be the maximum. And the most frequent type of strategy is the one where the input load is slightly higher than the penalty threshold.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：O225;D035

【相似文献】