群智能是近年来人工智能分析研究的一个热点话题蚁群算法作.docVIP

摘 要 群智能是近年来人工智能研究的一个热点话题。蚁群算法作为群智能算法的一个热点，是意大利学者M. Dorigo通过模拟蚁群觅食行为提出的。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的选择进行研究，得出其选择规律，并在以前学者“三步走”的基础上提出了一种“四步走”的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，同时以实验的方式对这几种算法各自求解TSP问题的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，最后还对陈烨的“蚁群算法实验室”的可视化编程进行了优化和改进，使之能够更方便的用于几种算法性能比较和同种算法不同参数的比较。 【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 Algorithm Xu Hui Abstract： Swarm intelligence has been a hot spot in the field of artificial intelligence in recent years. Among the algorithms of swarm intelligence, ant colony algorithm was presented by an Italy scholar M. Dorigo learning from 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. Key words： Swarm intelligence; Ant colony algorithm; Parameter Selection; TSP; Visualization 目 录 1 绪论 1 1.1 引言 1 1.2 群智能 1 1.3 蚁群算法的提出 2 1.4 本文研究工作 2 2 蚁群算法概述 4 2.1 蚁群算法基本原理 4 2.2 蚁群算法的优点与不足 5 2.3