A dynamic ant colony algorithm with disaster

Despite the numerous applications of ant colony algorithm in the optimization computation, there still exist, in the large-scale solution of problems, some shortcomings such as the long running time and the early convergence. This paper, on the basis of basic ant system algorithm, tries to improve the algorithm by means of the introduction of disaster, two-way search, 2-opt method, subsection save, dynamic innovation of such factor as pheromone. Some instances in TSPLIB have achieved or even exceeded the best solution ever recorded, which indicates that the improvement is excellent.