“Bees?” A Large Scale, Cooperative Simulation Weighing Altruism and Selfishness

In nature, communities that revolve around altruistic cooperation exist: for example, a bee colony, where individual bees retrieve nectar for the communal hive rather than scouting selfishly for their own survival. Although cooperation can certainly be implemented into an artificial life simulation, from an evolutionary standpoint a more interesting question is raised: can the interactions of a group and the parameters of their environment produce emergent cooperation among independent agents? This paper aims to investigate the possibility of emergent altruism through evolutionary game theory. To this effect, NEAT was used to evolve neural-net topologies across many generations in a bee colony simulation, testing what possible circumstances might lead to altruistic decisions or selfish decisions in the hive. A number of experiments were conducted, incorporating such concepts as a measure of group fitness, recurrent networks, and inter-agent communication. Initial speculation produced a hypothesis that the amount of altruism and selfishness demonstrated in NEAT-trained agents will be most affected by an individual fitness relying on group fitness. The results of the experiments demonstrate that while this is partly true and group fitness is key to altering artificial behavior, selfishness generally continues to prevail, with cooperation appearing rarely, or only when coerced. Through group-affected fitness, agents can learn to use altruism strategically to mitigate the negative effects of selfishness.