Game Mechanism Under Complex Topology Environment

Evolutionary game theory provides us with a set of concepts and methods, which can enrich our understanding of the world from a new point of view. Also,it well explains a number of seemingly contradictory but actually unifying phenomenon, such as cooperation and competition between the group. The complex network explain real network in an abstract way, which seem close to their essential characteristics, and it focuses on topological properties of the system environment. Because of it, it can deepen our understanding of the structure of complex systems. More importantly, it can provide a convenient system framework to decribe the relationship of individuals in the game. Thus, many studies begin to research on complex network combined with evolutionary game theory, which can further deepen the understanding of space game mechanism and reveal the nature of the evolution of cooperation among individuals. The paper is devoted to study the effect of the sparsity of research networks, the diversity of groups on the evolution of cooperation, so as to provide theoretical guide for promoting inter-group cooperation.The main content and innovative results of this paper are summarized as follows:(1) To discuss the level of cooperation of groups on sparse regular network,considering the influence of neighborhood size, population density, as well as collective selfishness on the cooperation rate of population. This research is mainly based on the classic game model : the Prisoner’s dilemma, the Snowdrift game, and the Stag-hunting game. The simulation results show that the cooperation level of groups has close relationship with the neighborhood, and typically there will be a moderate density value at which it can achieve the best cooperation level, namely threshold.(2) To propose a new kind of public goods game model with individual variety by introducing different factors of transferring strategy,we can simulate the decision-making process of the individuals. We can observe whether the effect on different neighborhood groups are the same. Studies have found the influential individuals joining the group has a significant promotion towards the level of cooperation.Moreover, the more various the learning strategy of two individuals, the more obvious the promotion.(3) Taking into account that a great number of snapshots are needed to further observe the distribution of cooperation and betray,we use the Java programming language under the Myeclipse to achieve the drawing function. Although the feature of the software is not much, but it is rather simple and easy to operate, which can save a lot of time in the process of the research.