Reputation preferences resolve social dilemmas in spatial multigames

Heterogeneity and difference in the dynamics of individual reputation may strongly affect learning behavior, and hence also the evolution of cooperation within a population. Motivated by this, we propose here an evolutionary spatial multigames model, wherein the reputation of an individual increases if they cooperate and decreases if they defect. After the payoffs are determined, individuals with a higher reputation will be more likely to act as strategy sources for other individuals. We perform systematic Monte Carlo simulations to determine the transitions between cooperation and defection, as well as the parameter regions of strategic coexistence. We show that preferential learning, based on dynamic reputation changes, strongly promotes cooperation regardless of the interaction network's structure. The mechanism responsible for more favorable evolutionary outcomes is enhanced network reciprocity, which leads to more compact cooperator clusters and thus to more robust spatiotemporal dynamics that are resilient to invading defectors. Our research may improve the understanding of selection patterns that favor the emergence and persistence of cooperative behavior. (paper)