A New Ant Population Based Improvement Heuristic for Solving Large Scale TSP.

Solving large-scale Traveling Salesman Problems (TSPs) has become a new challenge for the Ant Colony Optimization metaheuristic (ACO). Indeed, traditional ACO algorithms suffer from problems of stagnation and premature convergence and their computational cost increases remarkably as the problem size does. In this paper, we propose a new algorithm inspired by the ACO metaheuristic for symmetric large-scale TSPs. The basic idea of this new algorithm, called Ant-IH, is to change the classic role played by ants in order to accelerate the convergence speed. In fact, in our approach artificial ants are not anymore used to construct solutions but to improve the quality of a population of solutions using a guided local search strategy. To validate the performance of Ant-IH, we conduct the experiments on various TSPLIB instances. The experimental results show that our proposed algorithm is more efficient for solving symmetric large-scale TSPs than the compared algorithms.