摘要
蚁群算（Ant Colony Optimization，ACO）为了获得较优解，算法中的蚂蚁除了可以直接选择已经走过的路径外，也会选择未曾走过的路径，即“利用”与“探究”两种路径选择机制。这两种路径选择的概率是影响蚁群算法求解性能的关键。通过对旅行商问题（TSP）的仿真实验，结果表明，当“利用”被采用的概率很高时，可能会使蚁群算法的性能降低。当这两种机制被采用的概率差不多时，可以提高蚁群算法的性能。
In order to obtain a better solution, not only the ants in ACO （Ant Colony Optimization）can choose directly the path has been traversed directly, they can also choose the path has not been passed, that is ＂exploitation＂ and ＂explore＂ two kinds of path selection mechanism. The chosen probability of these two path selection mechanism may affect the performance of ACO. Through the simulation in TSP, the experimental results showed that when the ＂exploitation＂ a high probability of being adopted, it may cause performance degradation on ant colony algorithm. When these two kinds of mechanisms having been used in the probability are more or less, the performance of ant colony algorithm can be improved.
作者
唐晓寒
傅宏
张丽
TANG Xiao-han, FU Hong, ZHANG Li （1.College of Information Science and Engineering, Henan University of Technology, Zhengzhou 450001, China; 2. Zhengzhou Hualiang Technology Co., Ltd, Zhengzhou 450046, China）
出处
《电脑知识与技术》
2009年第10期7977-7978,共2页
Computer Knowledge and Technology
