Effects of Different Dynamics in an Ant Colony Optimization Algorithm

Finding shortest path in a labyrinth, made up of roads, crosses and dead ends, and where entrance and exit dynamically change during the time, is an important and challenging optimization task especially in emergency scenarios, such as earthquakes, volcanic eruptions, and/or hurricanes. In this research work we present a study on the effects of cooperative and competitive strategies in an agent-based model using an Ant Colony Optimization (ACO) algorithm for the solution of labyrinth problem. Two different ants' search strategies in the colony have been designed: those that acts competitively and selfishly, damaging some crossings (i.e. nodes) on the path, and cooperative ones, which instead attempt to repair them. The purpose of both strategies is finding a path from the entrance to the exit in order to gain the highest number of some resources positioned appropriately at the exit and bound to be collected if and only if both types of ants reach it via the shortest path. This research work has a two fold aim, that is, finding obviously the shortest path in the labyrinth (then maximize the resources gained), as well as analyzing the effects of both strategies on the overall ACO performances, and inspecting how one strategy affects the other by motivating it to improve its performances and its efficiency. From the overall outcomes, indeed, it emerges that the existence of the competitive ants is a strong incentive for cooperative ones to improve themselves.