Genetic Team Composition and Level of Selection in the Evolution of Cooperation

摘要: In cooperative multiagent systems, agents interact to solve tasks. Global dynamics of teams result from local agent interactions, and are complex difficult predict. Evolutionary computation has proven a promising approach the design such teams. The majority current studies use composed with identical control rules (ldquogenetically homogeneous teamsrdquo) select behavior at team level (ldquoteam-level selectionrdquo). Here we extend approaches include four combinations genetic composition selection. We compare performance genetically evolved individual-level selection, team-level heterogeneous simulated foraging task show that optimal combination depends on amount cooperation required by task. Accordingly, distinguish between three types tasks suggest guidelines for choice