Evolutionary Game Dynamics And Mechanisms Of Cooperation On Complex Networks

Emergence,referring to the existence or formation of collective behaviors,is one of fundamental features of complex systems.In particular,cooperation,which may emerge from various populations in nature and human societies,plays an important role in biological evolution and organizational hierarchies.However,cooperation is usually an altruistic behavior at the cost of self-interest,which is not consistent with Darwinian "survival of the fittest".Therefore,how to understand and explain the emergence of cooperation in selfish populations has long been a challenging issue.The intersection of evolutionary game theory and network science has given rise to a new branch of evolutionary games on complex networks.The complex networks describe the structure of population for interaction,the games depict the content of interactions between individuals,and the strategy updating rules endow individuals with dynamics.The evolutionary games on graphs can model a dynamical system that allows multiple individuals to interact intelligently.In this dissertation,we investigate the evolution of cooperation on complex networks based on evolutionary game framework and statistical physics.The main contents of this dissertation are as follows.First,we investigate the coevolutionary snowdrift game in a population,where multiple strategy update rules are employed,including aspiration,conformity,and imitation rules.The results show rich evolutionary transitions of the strategy update rules across the snowdrift T-S parameter space.One of the three strategy-updating rules prevails throughout the population in most parameter regions,while they could coexist in a small parameter region.Through the spatial snapshots,we find a spatially tessellated arrangement of strategies when the system is dominated by the aspiration update rule,the bistable absorption states pertaining to strategies,and the relationships between different strategies and strategy update rules.Meanwhile,we find that,in a large parameter area,the alliances of the conformity-driven and the aspiration-driven cooperators can boost the cooperation to a rather high level during the evolution.Next,we propose an adaptive strategy persistence mechanism and examine its impact on the evolution of cooperation in spatially structured populations.The mechanism determines the duration of strategy persistence,which is dynamically adjusted based on the comparison between individual payoffs and aspiration levels.We find that moderate aspiration values,which result in heterogeneous persistent time,consistently promote higher levels of cooperation.This is due to the distinct effects of the strategy persistence mechanism on cooperative and defective strategies in network settings.Further,considering that individuals who make decisions may be affected by environmental states,we propose an individual strategy persistence mechanism based on local and global environmental comparisons in the evolutionary prisoner’s dilemma game.We find that environmental feedback can effectively promote cooperation,especially when the frequency of cooperators is low.By comparing with the static or stochastic strategy-persistence scenarios,we note that the strategy persistence based on environment comparison is more effective in promoting cooperation,especially at low-noise situations.The main conclusions are robust to network structures.Finally,based on the environment comparison,we further propose an individual migration mechanism on complex networks,where individuals decide whether to migrate or not based on the comparison between local and global environments.Through an analysis of the evolutionary process of the system,we observe that cooperator clusters expand during evolution and merge to form larger clusters through migration,ultimately leading to increased frequency of cooperators in the population.Furthermore,our findings suggest that individual rationality plays a more crucial role in the decision-making process related to migration than in strategy learning.In this dissertation,we explore various mechanisms for the promotion of cooperation on complex networks under the framework of evolutionary game theory.Multiple evolutionary game models that promote cooperation have been proposed,including the co-evolution of multiple strategy update rules,adaptive strategy persistence mechanisms driven by individual aspiration or environmental feedback,and individual migration based on environmental feedback.Through a thorough investigation of the evolution of cooperation within the contexts of the prisoner’s dilemma game,snowdrift game,and static or dynamic interaction structures,this study provides valuable insights into biological evolution,the emergence of cooperation,and the mitigation of social dilemmas.These findings bear significant implications for both social and biological systems.