Abstract: To overcome the drawbacks of the conventional ant colony algorithm such as slow convergence and low precision, a returned ant algorithm is proposed. To improve global convergence, an outside force makes ants move according to the whole city’s distribution. Moreover, by comparing the distances between the current position and the others, computational complexity of regional optimization is reduced. The simulation results of several traveling salesman problems show that the proposed algorithm is feasible and highly efficient.

Key words: ant colony system(ACS), global convergence, traveling salesman problem(TSP)