A Computational Analysis on a Hybrid Approach: Quick-and-dirty ant colony optimization

Ant colony optimization (ACO) is a well known metaheuristic. In the literature it has been used for tackling many optimization problems. Often, ACO is hybridized with a local search procedure. All the solutions generated by ants (or some of them) are improved by the local search. In this paper we propose a different framework, called Quick-and-dirty ant colony optimization. It is an hybrid approach based on the sequential coupling of ACO and a local search. It exploits the ability of ants to explore the search space. After ants point out the most promising area, the local search procedure is used for analyzing it. Computational experiments on the traveling salesman problem confirm that when the search space is large, by allowing the local search to concentrate on a smaller region it is possible to improve the quality of the performance, provided that this region is properly selected.