Parameters Research and Improvement of Ant Colony Algorithm

摘要

The ACA (Ant Colony Algorithm) parameters a, β, p, q0 and so on have a great effect on algorithm performance. The optimal solutions in different configuration parameters are acquired through TSP simulation experiment. In accordance with experimental data, the paper discusses effect of different configuration parameters on results and analyzes the causes as well as investigates the relation between parameters and algorithm performance. On the basis of modified ACA and through updating the rules of global pheromone, ants can concentrate on better routes and thus convergence rate is improved.