The Emergence Pattern And Evolution Of Altruistic Behavior On Complex Networks

In the context of rapid and changeable development of society, competitive environmenthas led to the prevalence of egoism. Food and drug safety issues, chain swindle fromenterprise vendors, corruption of government officials and moral deficiency of youths hascaused serious social harm and crisis can’t be ignored. The research on altruistic behaviorhelps to solve practical problems exists in social construction. An urgent need to the goodsociety atmosphere triggered widespread concern of scholars in various fields. How thealtruistic behavior emerges and maintains? What factors can influence and restrict altruisticbehavior? The two problems have already become hot research topics. Meanwhile, gametheory provides an important means of modeling and analyzing altruistic behavior.Ubiquitous networks in nature and human society are complex networks. Thedevelopment of complex network theory provides a framework widely applied to describegame relationship. The points in network represent individuals, and the links between pointsrepresent relationship among individuals. The topological structure of network and game rulesboth impact on individual’s behavior. Microscopic interactions between individuals enablemacroscopic phenomena in group. Game on complex networks pays close attention to thecooperation emerged among selfish individuals, and studies the mechanism on spontaneousaltruistic behavior by way of establishing certain rules. It provides a theoretical basis forunderstanding altruistic behavior and exploring mechanism promoting cooperation.In order to study the emergence pattern and evolution of altruistic behavior in depth, weproposed a mechanism based on reputation and future expectation in the framework of gametheory and complex network theory. Two structures are considered, including spatialsmall-world network model and scale-free network model. Each player interacts with hisimmediate neighbors can follow two strategies: cooperation or defection. Reputation as localinformation is used to measure and evaluate the historical behavior of individuals. Both sidesin game can know the historical behavior each other before making their decision.Individual’s reputation can be used to predict the probability of choosing cooperation byopponent. The individual with high reputation is more easily to be chosen. In addition,discount factor is introduced to describe the future expectation of players. For the purpose ofpayoff-maximizing in repeated games, the higher the future expectation of player is, thebigger the probability of choosing cooperation. Hence, players not only consider currentpayoff but also care about future payoff when they employ strategy. Analysis and simulations show players choose cooperation voluntarily for the purpose ofpayoff-maximizing as long as the benefit-to-cost ratio is big enough on the two networks. Thelevel of convergence average reputation of group does not absolutely rely on the level ofinitial average reputation of group. Group can converge to full cooperation even though theinitial average reputation of group is small. In a certain interval, a slight increase of the initialaverage reputation of group can effectively enhance the level of cooperation. We foundcomplex networks have a certain ability to resist the disturbance by simulating thenon-subjective factors which prevent the cooperation from emerging. The cooperation state inscale-free network is more stable than that in small-world network.Concentrating on the situation of group state changes from cooperative to defective,dynamic spatial patterns of small-world network show long-range connections are the mainreason for the emergence of new defective cluster. Reasonable randomness of network is infavor of motivating altruistic behavior. In scale-free network, heterogeneity makes playersconverge to three states: full cooperation, full defection and reputation oscillate in a specificdomain with particular mode. A small amount of players with large degree and highreputation and a large number of players with small degree and low reputation restrict eachother, which enable group can converge to a state between full cooperation and full defection.In addition, heterogeneity has no effect on promoting behavior spread under the mechanismagainst the common belief.