The Research On Multi-person Game And Evolution Of Cooperation

Cooperation is ubiquitous both in biological and social system, researching the emergence and the evolution of cooperative behaviors is one of the most important issues in the evolutionary biological and social theory. Game theory, especially the evolutionary game theory, provides an effective theoretical framework for the scientists to study the issue. Given the complexity of the actual social network as well as the complicated interaction forms of the individuals, researches of game theory on the complex network have a rapid development recently.In this thesis, we overview the development of evolutionary game theory, especially evolutionary game theory on complex network. We elaborate the emergence of cooperation and evolutionary dynamics from the evolutionary models, topology of network, game evolving rules and so on. Considering the fact of the diversity of interaction among the individuals, we investigate the emergence and evolution of cooperation in spatial multi-person games and its applications in wireless network. The works are as follows:1. Based on the traditional snowdrift game, we investigate the evolution of cooperation in multi-person snowdrift game on two different networks (well-mixed network and regular lattice network). On well-mixed networks, the research shows that the cooperation frequency decreases as cost-to-benefit ratio increases. The experiment also shows that the more individuals involved in the game, the more rapid cooperation frequency drops. This indicates that individuals tend to wait for someone else to perform the task and enjoy a free ride. In comparision with well-mixed network, the regular lattice is deteriorative to cooperation for multi-person snowdrift game. It is because that cooperates are always explored by neighbour defectors. As a result, cooperation dies out eventually.2. We investigate the spatial multi-person snowdrift game with critical mass M. In stark contrast to the situation that traditional spatial multi-person snowdrift game inhibits cooperation, cooperation can be greatly enhanced when M is at a certain range. For a given cost-to-benefit ratio r, there is an optimal critical mass Mop to reach the optimal cooperation level and there are discontinuous phase transitions and steplike behaviour when M increases. When the number of players involving in the game increases, the steplike structures also increase and the optimal cooperation fop frequency shifts to larger M. These nonlinear features can be explained by constructing the payoff level structures.3. We investigate the applications of multi-person evolutionary games in wireless networks. In specific applications, we propose a packets forwarding evolutionary game model in wireless network by introducing a reciprocity strategy besides cooperation and defection strategy. We theoretically analyse the dynamics mechanism and stability of the model, which is validated by the simulation. In experiments, we discuss the evolutionary mechanism of the system in different initial distribution ratio, reciprocity cost and mutation probability. By providing the reciprocity strategy, the original selfish nodes would shift to cooperation in certain conditions, as a result, the packet forwarding service of the network can be fulfilled.