Active Inferants: An Active Inference Framework for Ant Colony Behavior

In this paper, we introduce an active inference model of ant colony foraging behavior, and implement the model in a series of in silico experiments. Active inference is a multiscale approach to behavioral modeling being applied across settings in theoretical biology and ethology. We specify and simulate a Markov decision process (MDP) model for ant colony foraging using the alternating T-maze...