Bacterial Evolution Algorithm for rapid adaptation

Abstract

In this paper, we propose a Bacterial Evolution Algorithm (BEA), inspired by the mechanism of bacteria rapidly adapting themselves to an ever-changing environment. In this paper, we call adaptive agents bacteroids. Bacteroids have their own fitness function that reflects the rates of energy replenishment and collision avoidance. The characteristic of this algorithm is that a selection of bacteroids is made by their environment with their death as a trigger of the selection. This selection is, in general, done irrespective of the bacteroids own fitness function. Even if they have a higher fitness value at a given moment, they will die when they are exposed to a sudden severe environmental condition. If some bacteroids die, the strongest adjacent bacteroids will take over their bodies by inserting their chromosomes, which should be most adaptive in that local area. Mutation is applied at the moment of this takeover to give the bacteroids a chance at evolution. The BEA is appropriate in an environment where many agents like bacteroids are working together in one place with many chances of interaction.