Ant Colony Optimization And Its Application

Discrete optimization plays a great role both in thoery and application. It's almost impossible to retain the optimal solution with deterministic algorithm especially when the scale of the problem is very large. Ant colony optimization(ACO) belongs to meta-heristic algorithms and it can help us to obtain a reasonable optimal solution. ACO has such features as positive feedback, parallel computing,robustness and so on.Any discrete optimization can be settled by ACO if modified properly. Great breakthrough has been made in performance of ACO since the appearance of ant system. Its application includes discrete optimization, net routing, machine learning, image processing and so on.This paper concentrates on the principle of ACO, improved versions of ant system,the theory, as well as stategies to improve the performance of the algorithm. The discipline and the main steps to apply ACO are also discussed in the paper. Besides, some instances are given to illustrate the great performance of ACO. The main contribution comes as follows:An improved Ant Colony System (IACS), which employed a new factor in transition rule to overcome the premature behavior in (ACO) is proposed. The factor can help the ants to obtain a better result by exploring the arc with low pheromone trail accumulated so far as time elapses. Besides, it can avoid the over-concentration of pheromone trail to enlarge the searching range. Simulation results shows that the IACS has better performance in solving the Traveling Salesman Problems (TSP) and more outstanding global optimization properties.