信息素增量动态更新的改进蚁群算法  

Ant Colony Algorithm with Dynamic Pheromones Increment Updating

作　　者：贾瑞玉[1] 张新建[2] 冯伦阔[1] 李永顺[1] 

机构地区：[1]安徽大学计算智能与信号处理教育部重点实验室,安徽合肥230039 [2]安徽大学计算机科学与技术学院,安徽合肥230039

出　　处：《计算机技术与发展》2009年第9期32-34,37,共4页Computer Technology and Development

基　　金：安徽省自然科学研究项目(kj2008B092)

摘　　要：蚁群算法是一种优秀的拟生态启发式算法,具有较强的鲁棒性,易与其它拟生态算法结合等特点。不过,它也存在着容易陷入局部最优、收敛速度慢等问题。通过分析蚁群算法的运行机制,得出了蚁群算法易陷入局部最优、收敛速度慢等不足产生的原因,针对这些不足,通过修改基本蚁群算法中信息素的更新规则,使得每轮搜索后信息素的增量能更好地反映解的质量,以加快收敛;同时引进了路径信息素平滑机制以平衡路径上的信息量。对TSP问题的仿真实验结果表明,改进后的算法加快了收敛速度,提高了全局搜索能力。Besic 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.

关 键 词：蚁群算法 信息素更新规则 全局收敛 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]