Contribution diversity and incremental learning promote cooperation in public goods games

Understanding the evolution of cooperation in nature has long been a challenge and how to promote cooperation in public goods games (PGG) has attracted lots of attention recently. Social diversity has been found helpful to explain the emergence of cooperation in the absence of reputation and punishment. However, further refinement on how individuals reallocate their contribution to each PGG remains an open question. Moreover, individuals in existing works mostly teach or learn from neighbors according to their payoff in the last generation only. However, individuals in reality are preferred to learn from others with a long-term good performance. Therefore, in this paper, a new contribution diversity (CD) is designed and incremental learning (IL) is introduced. We investigate how these two may influence the evolution of cooperation in PGG. Based on the simulation results, we found that both the CD and IL can promote the cooperation in PGGs. Moreover, when cooperators are shaken in their strategy, CD may fail in reallocating contribution of individuals properly. However, IL is found effective to stabilize faith of cooperators and cooperators under IL reflect a long-term advantage over defectors in terms of benefits. Therefore, we further find IL and CD can mutually benefit each other in promoting cooperation, as CD can reasonably adjust the investment of cooperators while IL can provide more information to CD.