A Co-Evolutionary Hybrid ACO for Solving Traveling Salesman Problem

Abstract

Ant Colony Optimization (ACO) is an approximate method proposed recently. Many ACO based approaches and hybrid methods have been proposed for solving the traveling salesman problem (TSP); However, the balance between intensification and diversification is also difficult to solve. In this paper, we propose a co-evolutionary hybrid method (CEACO-GA) by adopting multiple colonies which perform ACO or GA algorithms, and a co-evolutionary strategy among colonies which is to enhance the interaction among colonies by communication between ACO and GA colonies, thereby to control the population diversity. The number of colonies that perform GA operations is used to adjust the balance between intensification and diversification. The CEACO-GA is tested on various problem instances in the TSPLIB standard library, and the results of numerical calculation show that the the CEACO-GA has more outstanding performance comparing to other algorithms.