Evolutionary dynamics of multiplayer cooperation under mobile structured populations #

Diogo L Pires, Erovenko, Igor V, Broom, Mark

12:10 Monday in 3Q16.

Part of the Game theory and agent-based models session.

Abstract #

Evolutionary models are used to study the self-organization of collective behaviour, often incorporating population structure due to its ubiquitous presence and long-known impact on emerging phenomena. We investigate realistic multiplayer interactions by considering interacting groups that arise from the encounters of individuals moving on networks contingent on who they meet. In this context, we find that the evolution of multiplayer cooperation is primarily dependent upon the network topology, while different stochastic update rules seldom influence evolutionary outcomes. Cooperation robustly co-evolves with movement on complete networks, while structured topologies have a detrimental effect. These findings contradict the established rule in evolutionary graph theory that cooperation can only emerge under some update rules, and if the average degree is low and topology far from complete. We believe that group-dependent movement erases the locality of interactions, suppresses the impact of evolutionary structural viscosity on the fitness of individuals, and leads to assortative behaviour that is much more powerful than viscosity in promoting cooperation. These factors equate the impact of population structure on the evolution of cooperation with its ability to sustain assortative movement between cooperators, while hindering the exceptional significance of some update rules. We analyze the minor lasting differences between update rules through a comparative analysis of evolutionary outcomes and fixation probabilities.