Evolutionary Games And Mechanisms Of Cooperation Evolution On Complex Networks

The phenomena of cooperation is widespread in nature and human society.How to understand the emergence and evolution of cooperation among unrelated selfish individuals is a major scientific puzzle.As a significant extension of the traditional evolutionary game theory,evolutionary games on graphs provides an important framework to understand and analyze the decision-making behavior of individuals and groups under complex interactive environment,especially the occurrence and evolution of cooperative behavior.In previous works,the investigation is primarily from a perspective of rational economic individuals,the adoption of game strategy and the adjustment of game relationships are mainly depend on comparison result of payoff.Nevertheless,the effects of other impact factors such as reputational preference,other-regarding preference,and trust,etc.on players’ decision-making behavior are ignored.Hence,our focus is to integrate other factors impacting real individuals’ decision-making behavior that have been found in experimental economics and behavioral science into the evolutionary games on complex networks.Moreover,we take the coevolution and interactive effects of structural dynamics,link weight dynamics and strategy dynamics into consideration,and has carried out the following works:(1)The behavioral observability and reputational preference are introduced into the spatial prisoner’s dilemma game(PDG).A dynamically changing reward is established for cooperative neighbors whose reputation is higher than the average score of all neighbors.The reward is determined by the observability of game behavior and the influence strength of reputational preference.The effects of rewarding mechanism on cooperation evolution and the evolutionary characteristics of cooperation in different network topologies are studied.Simulation results show that,compared to the traditional PDG,the rewarding mechanism significantly promotes the emergence and evolution of cooperation in homogeneous,heterogeneous,and real social networks.Under the mutual impacts of behavioral observability and reputational preference,the evolutionary advantage of cooperators is extremely enhanced.Compared with regular square lattice,it is easier for cooperation to emerge and maintain in WS small-world network and ER random network.(2)The other-regarding preference impacting decision-making behaviors of real individuals is integrated into the traditional prisoner’s dilemma game.Under the influence of other-regarding preference,players are willing to sacrifice personal payoff to reward the cooperative neighbors who made significant contribution to the mutual cooperation.The effects of other-regarding preference based reciprocal mechanism on the evolution of cooperation,and the characteristics of cooperation emergence and spread in four different networks are investigated.Simulations results reveal that the reciprocal mechanism has a positive impact on the occurrence and evolution of cooperation in regular square lattice,WS small-world network,ER random network,and American college football network.Both the dynamic adjustment of link weight and the increase of influence strength of other-regarding preference play an essential role in the emergence of cooperative behavior.In the above four networks,the ER random network mostly favors the occurrence and sustainability of cooperation.The high clustering of WS small-world network and regular square lattice accelerates the evolution of cooperation.(3)A coevolutionary model of network structure and game strategy is constructed by incorporating behavioral punishment-based link dismissing rule and reputational-preferencebased edge establishment rule into the traditional prisoner’s dilemma game.The influence of the terminating and reconnection mechanism on the evolution of cooperation as well as the underlying game interaction network structure are explored in homogeneous regular square lattice.Simulation results suggest that the dynamic coevolution of network structure and game strategy fosters the emergence of an underlying heterogeneous interactive network,which provides a favorable network topology for the occurrence and evolution of cooperation under strategy dynamics.With the increase of temptation to defect and the density of game interaction network,the adjustment frequency of network structure needs to be increased to support the emergence and spread of cooperation.Under the impact of coevolutionary model,cooperators gradually occupy the vertices of high degree;the initially homogeneous regular square lattice finally evolved as a scale-free network with disassortative,which displays high heterogeneity with a power-law degree distribution.(4)The trust existed in real social interactions is introduced into the spatial prisoner’s dilemma game.A dynamically changing incentive reward is established for the cooperative neighbor who has the greatest link weight.The effects of trust-based incentive mechanism on the evolution of cooperation in PDG,as well as the impacts of distinct game interaction networks on cooperation evolution are investigated.Simulation results show that the incentive mechanism has a significant role in the promotion of emergence and spread of cooperation in regular square lattice,WS small-world network,ER random network,and American college football network.The adaptive adjustment of cooperative trust with the dynamic change of game strategy favors the evolution of cooperation.From the influence of network structure on cooperation evolution perspective,the small-world effect inhibits the emergence of defection;the high heterogeneity of network facilitates the emergence and survivability of cooperation;the high clustering of network accelerates the evolution of cooperation.Findings of the incentive mechanisms of cooperation not only provide references for the solution of social dilemmas,especially the establishment of cooperation promotion mechanisms in human social society,but also propose references for enhancing the effective cooperation between multi-agents and the autonomous coordination capabilities of multi-agent systems,which are significant for alleviating the urban traffic congestion.