A Preliminary Study for Ant Colony Optimization with a new Reinforcement Strategy

Abstract

Different Ants Colony Optimization (ACO) algorithms use pheromone information differently in an attempt to improve their relative performance. In this paper, we describe a new systematic reinforcement strategy as a means to improve the pheromone update rules of existing ACO algorithms. We examine the proposed strategy and compare it with other improvement strategies using the well - known Traveling Salesman Problem (TSP). The results indicate that the performance of both the AntSystem (AS) and the Ant Colony System (ACS) algorithms is improved by applying the proposed strategy. We postulate that the proposed strategy allows the ants, in some sense, to both refine the search in promising regions, and escape explored areas of the search space more consistently and effectively than other reinforcement strategies.