Ant Colony Algorithm with Dynamic Pheromones Increment Updating

贾瑞玉,张新建,冯伦阔,李永顺

DOI: https://doi.org/10.3969/j.issn.1673-629x.2009.09.009

2009-01-01

Abstract:Basic ant colony is an excellent meta-heuristic algorithm and has strong robustness,easy to mix together with other meta-heuristic algorithm,but it also has slow convergence and easily be trapped in a local optimum.Could get the reasons that the ant colony has slow convergence and easily be trapped in a local optimum from analying the principle of the algorithm and aiming at these disadvantages,present a new solution that modifies the rule of updating pheromones in ACA,so that after every round of search,the increment of pheromones can better reflect the quality of a solution to quicken the convergence;also use the solution that pheromones smoothing to balance the pheromones on the route.Experimental results for solving TSP problems indicate that the improved algorithm increases the speed of convergence and enhances the ability of searching the whole best solution.