群智能是近年来人工智能研究分析的一个热点话题蚁群算法作.docVIP

个人收集整理 仅供参考学习 个人收集整理 仅供参考学习 PAGE / NUMPAGES 个人收集整理 仅供参考学习 摘 要 群智能是近年来人工智能研究地一个热点话题.蚁群算法作为群智能算法地一个热点，是意大利学者M. Dorigo通过模拟蚁群觅食行为提出地.本文首先介绍了群智能，然后详细介绍蚁群算法原理及其优缺点.接着依据大量实验对参数m、α、β、ρ、Q地选择进行研究，得出其选择规律，并在以前学者“三步走”地基础上提出了一种“四步走”地有效方法来选择蚁群算法最优组合参数，然后对蚁群改进算法进行分析，同时以实验地方式对这几种算法各自求解TSP问题地性能进行了对比分析，得出性能结果排名，并发现当TSP问题最优解相同时还可以依据其他性能（迭代次数、迭代时间等）得出最优结果，最后还对陈烨地“蚁群算法实验室”地可视化编程进行了优化和改进，使之能够更方便地用于几种算法性能比较和同种算法不同参数地比较. 【关键词】 群智能；蚁群算法；参数选择；TSP；可视化 Experimental Analysis and Parameter Selection for the Ant Colony Optimization Algorithmb5E2RGbCAP Xu Hui Abstract： Swarm intelligence has been a hot spot in the field of artificial intelligence in recent years.Among the algorithms of swarmintelligence,ant colony algorithm was presented by an Italy scholar M.Dorigo learningfrom the behavior simulating ant colony foraging. Firstly,this paper has introduced the group intelligence and promoted the ant colony algorithm,obtained the choice regular of “m,α,β,ρ,Q” basing on the experiment, and proposed an effective method named“four steps” in the fundation of others scholars’ “three steps” to choose the most superior combination parameter of ant group algorithm,then analied the six kinds improved algorithm of ant colony ant colony algorithm, at the same time explained the ability of several kinds of ant algorithm to solve the TSP question according to the experiments; obtained the most superior result according to other performance (iterative number of times, iterative time and so on) when the most superior result of TSP question optimal solution is same. Finally, this paper also has carried on the optimization and the improvement to the visible programming of ChenYe’s“ant colony algorithm laboratory” to enable it more convenient to use in several algorithms performance comparison and the comparison of different parameter and homogeneous algorithm.p1EanqFDPw Key words： Swarm intelligence; Ant colony algorithm; Parameter Selection; TSP; VisualizationDXDiTa9E3d 目录 TOC \o 1-3 \h \z \u HYPERLINK \l