Evolution of Attraction for Cooperation

We are investigating an attraction mechanism that models interpersonal relations such that an individual acts differently, and presumably more cooperatively, with those to whom it is attracted than with others. This is a generic mechanism and the types of relations envisaged can range from regular physical attraction to attachment of toddlers to caregivers. We have shown in the past that the introduction of this attraction mechanism in an iterated prisoner's dilemma game yields higher average agent scores in tournaments within uniform or mixed populations. In the present work, we show that this mechanism has an evolutionary advantage and that it can evolve in various ways (probability of attraction, network of attraction). These results show that cooperation and social stability can be enhanced by psychological mechanisms that are external to the game setting but interfere with it and that these mechanisms may be further selected and reinforced by evolution and catalyze cooperation. We further show that evolution of attraction in the context of cooperation is possible in fairly small networks of homogeneous agents with noise but not in networks of heterogeneous agents without noise. This remark has interesting implications for the study of the origins of human cognition.