Abstract:Due to the shortcomings of ant colony optimization(ACO) algorithm, such as premature convergence and exorbitantly long computation time, an enhanced ACO algorithm is proposed. It constructs a good solution pool which holds a certain number of best solutions found so far. These solutions are clustered by a developed neighbourhood-based clustering algorithm, and accordingly some different regions which contain good solutions can be captured. The proposed ACO algorithm alternately employs the good solutions belonging to different clusters to update pheromone. By this means, both the intensification and the diversification of search are consulted. Simulation experiment is conducted on typical travelling salesman problems. The results show that, the presented algorithm is more efficient in generating high-quality solutions.