蚁群算法在求解TSP问题中的改进研究

针对蚁群算法在求解大规模优化问题时存在的3个缺点：消耗时间长、蚂蚁在下次搜索时目标导向不强导致搜索随机性大、寻优路径上的信息素过度增强导致得到假的最优解。本文提出了基于边缘初始化和自适应全局信息素的改进蚁群算法。在相同参数下,其搜索时间大大缩短,并且得到了更好的最优解。将其应用到旅行商（TSP）问题中,和基本蚁群算法、遗传算法相比较,其具有以下优点：较好的搜索最优解的能力;对新解不会过早的终止;探索新解的能力进一步增强。因此,改进的蚁群算法在求解TSP等组合优化问题时非常有效。

The improvement of ant colony algorithm in solving TSP

To cope with the three drawbacks of ant colony algorithm in solving large -scale optimization problems： long time- consuming ,the big randomness of the ant＇s search due to the weakness of the goal-orientation,the false optimal solution due to the excessive enhancement of the pheromones of the optimal path,the improved ant colony algorithm which is based on the edge initialization and the adaptive global pheromone is presented in the paper.Under the same parameters,its searching time shortens greatly and it gets a better optimal solution.Apply it to the traveling salesman problem,compare it with the other two methods which are basic ant colony algorithm and GA,it has the following advantages：it has a better capability in searching optimal solution;it can＇t be prematurely terminated to the new solution;the ability of exploring new solutions is enhanced.So,the improved ant colony algorithm is very effective in solving the combinatorial optimization problems such as TSP.