基于随机交互强度机制的空间博弈的仿真研究

发布时间：2018-06-12 00:10

  本文选题：合作演化 + 囚徒博弈　； 参考：《计算机仿真》2017年08期

【摘要】：空间博弈研究中多假定网络上物理相邻的个体必然发生博弈,忽略了真实系统中个体与邻居发生交互不具确定性。从合作演化的角度,利用空间囚徒博弈理论,依据策略性质对博弈分类,提出一种"随机交互强度"机制,使博弈以固定概率随机地发生在相邻个体间,建立了基于二维正方晶格网络的博弈演化模型,并使用多主体系统进行了仿真。利用系统平均合作水平作为判据,对比连续策略系统与离散策略系统的不同交互强度,给出了仿真结果。结果表明,个体间博弈的随机性可以提高离散策略系统的平均合作水平,较低的交互强度可使系统达到合作稳定态,相比于必然性交互机制,上述机制可以有效促进持离散策略个体间的合作。
[Abstract]:In the study of spatial game, it is assumed that the physical adjacent individuals on the network must play games, and the interaction between individuals and neighbors in real systems is not deterministic. From the point of view of cooperative evolution, using the theory of spatial prisoner game and classifying the game according to the nature of strategy, a mechanism of "random interaction intensity" is proposed, which makes the game happen randomly between adjacent individuals with fixed probability. A game evolution model based on two-dimensional square lattice network is established and simulated with multi-agent system. The average cooperation level of the system is used as the criterion to compare the different interaction intensity between the continuous policy system and the discrete policy system, and the simulation results are given. The results show that the randomness of the game between individuals can improve the average cooperation level of the discrete strategy system, and the low interaction intensity can make the system achieve cooperative stability, compared with the inevitable interaction mechanism. The above mechanism can effectively promote the cooperation among individuals with discrete strategy.
【作者单位】： 大连理工大学软件学院;
【基金】：国家自然科学基金资助课题(61300087) 辽宁省自然科学基金(2015020003) 中央高校基本科研业务费专项资金资助(DUT15QY40)
【分类号】：O225
，

本文编号：2007305