The Evolution Of Cooperation Based On Individual Strategies And Migration Mechanisms

Since the beginning of human society,cooperation has been a major issue affecting the development of human being.Individuals take action under the motive of self-interest,and Individual behavior does not directly lead to collective common interests.This raises the study of how cooperation is generated and evolved in groups of selfish individuals.Game theory provides a strong theoretical framework for the study of evolution of cooperation.Nowak has summarized five fundamental cooperation mechanisms,including direct reciprocity and spatial reciprocity.Besides the basic mechanisms,reputation,migration,and teaching activities have been widely reached.So far,the issues of strategy updating mechanism,environmental factors and migration mechanism are still not clear and need to be further explored.However,under different network structures and game models,various mechanisms provide alternative information to agents,which will significantly affect the emergence of social norms.Therefore,in this thesis,we focus on three specific scenarios:iterated prisoner's dilemma game with noise,prisoner's dilemma game on static networks and prisoner's dilemma game on mobile networks.This thesis proposes a flexible strategy,strategy updating rules and migration mechanism for cooperation emergence in each scenario and analyzed emergence process of each mechanism.The main contributions of this dissertation are listed as follows:First of all,this thesis focuses on the repeated prisoner's dilemma game under the influence of noise.This thesis proposes a flexible strategy with reciprocal altruism(FRAM).Based on FRAM,the player has a certain degree of tolerance when opponent's strategy is defection.FRAM makes player be able to keep long-term relationship with another player with FRAM.The simulation results and microscopic analysis show that FRAM and CTFT have strong anti-noise performance,and the TFT is easy to win in low noise environment.We also find that the higher interaction frequency between individuals,the better performance of cooperative strategies,such as FRAM and CTFT.On the contrary,gradually loses its advantage with decreasing interactive frequency.The research shows that the strategy based on individual long-term memory has strong ability to resist noise,and shows greater advantages in the environment where individuals interact frequently.It also shows that environment has great influence on game strategy.Secondly,this thesis focuses on the scenario of static spatial games.A gradual strategy learning mechanism based on the income difference between adjacent individuals is proposed.Based on the mechanism proposed,if the income difference between adjacent individuals is large,the individual will imitate the strategy of the adjacent individual greatly.The simulation results and microscopic analysis show that the gradual strategy learning mechanism can promote cooperation when the parameter of sensitivity to income differences is small.Compare with the strategy take-over mechanisms that most previous works adopted,the mechanism proposed in this thesis is more conducive to cooperation,especially when individuals know fewer information of neighbors' income.The research shows that individuals with middle strategy form a disjunctive band between cooperators and defectors during strategy evolution process.Thirdly,this thesis focuses on the scenario of dynamic spatial games.A coevolutionary mechanism based on individual evolutionary preference is proposed.Based on the mechanism proposed,individuals have two types of preferences:imitation and migration.The simulation results and microscopic analysis show that the larger migration preference or strategy imitation preference can both achieve the explosive expansion of cooperation,and no tendency is not conducive to cooperation.Based on previous works,migration can promote cooperation because it can adjust the border of the cooperator clusters.The research shows that migration can help cooperators find each other and form new clusters.The research also proved that the latter way is more conducive to cooperation in low population density environment.Finally,this thesis focuses the problem that the defectors form local clusters based on payff-driven migragion.Two mixed-mode mechanisms based on payff-driven migragion is proposed.PDMMP mechanism combines payoff-driven migration and global-mean-payoff-driven and migration;PDME mechanism combines payoff-driven migration and escaping-from-defectors migration.PDMR which combines payoff-driven migration and random migration is used for comparison.The simulation results show that When the desity of population is low,payoff-driven migration mixing a fewer other modes of migaration can promote cooperation dramatically.When the desity of population is high,payoff-driven migration mixing a high rate of escaping-from-defectors migration or mediate rate of global-mean-payoff-driven can promote cooperation significantly.The research shows that PDMMP and PDME can eliminate the clusters of defectors effectively and improve the level of cooperation significantly compared with PDMR.