Evolutionary optimization of pheromone-based stigmergic communication

Abstract

Pheromone-based stigmergic communication is well suited for the coordination of swarm of robots in the exploration of unknown areas. We introduce a guided probabilistic exploration of an unknown environment by combining random movement and stigmergic guidance. Pheromone-based stigmergic communication among simple entities features various complexities that have significant effects on the overall swarm coordination, but are poorly understood. We propose a genetic algorithm for the optimization of parameters related to pheromone-based stigmergic communication. As a result, we achieve human-competitive tuning and obtain a better understanding of these parameters. © 2012 Springer-Verlag.