Learning What to Eat: Studying Inter-relations Between Learning, Grouping, and Environmental Conditions in an Artificial World

In this paper we develop an artificial world model to investigate how environmental conditions affect opportunities for learning. We model grouping entities that learn what to eat in a 2D environment. We study diet development and focus on the social consequences of individual learning in relation to different environmental conditions. We find that homogeneous and patchy environments have opposite effects on learning. Homogeneous environments lead to diet differentiation, while patchy environments lead to diet homogenization among the members of a group. In patchy environments, grouping results in a social influence on individual learning and could be the simplest way to achieve social inheritance of information. Moreover, diet differentiation can affect group cohesion, leading to group fragmentation along dietary lines. This suggests that if social learning leads to diet homogenization, it could play a role in maintaining group cohesion.