Swarm communication by evolutionary algorithms

This research has applied evolutionary algorithms to evolve swarm communication. Controllers were evolved for colonies of artificial simulated ants during a food foriaging task which communicate using pheromone. Neuroevolution enables both weights and the topology of the artificial neural networks to be optimized for food foriaging. The developed model results in evolution of ants which communicate using pheromone trails. The ants successfully collect and return food to the nest. The controller has evolved to adjust the strength of pheromone which provides a signal to guide the direction of other ants in the colony by hill climbing strategy. A single ANN controller for ant direction successfully evolved which exhibits many separate skills including food search, pheromone following, food collection and retrieval to the nest.