Dynamics of heuristics selection for cooperative behaviour

Situations involving cooperative behaviour are widespread among animals and humans alike. Game theory and evolutionary dynamics have provided the theoretical and computational grounds to understand what are the mechanisms that allow for such cooperation. Studies in this area usually take into consideration different behavioural strategies and investigate how they can be fixed in the population under evolving rules. However, how those strategies emerged from basic evolutionary mechanisms continues to be not fully understood. To address this issue, here we study the emergence of cooperative strategies through a model of heuristics selection based on evolutionary algorithms. In the proposed model, agents interact with other players according to a heuristic specified by their genetic code and reproduce -- at a longer time scale -- proportionally to their fitness. We show that the system can evolve to cooperative regimes for low mutation rates through heuristics selection while increasing the mutation decreases the level of cooperation. Our analysis of possible strategies shows that reciprocity and punishment are the main ingredients for cooperation to emerge, being conditional cooperation the more frequent strategy. Additionally, we show that if in addition to behavioural rules, genetic relatedness is included, then kinship plays a relevant role. Our results illustrate that our evolutionary heuristics model is a generic and powerful tool to study the evolution of cooperative behaviour.