当前位置:首页 > IT计算机/网络 > AI人工智能 > 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...
I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IIExperimentalAnalysisandParameterSelectionfortheAntColonyOptimizationAlgorithmXuHuiAbstract:Swarmintelligencehasbeenahotspotinthefieldofartificialintelligenceinrecentyears.Amongthealgorithmsofswarmintelligence,antcolonyalgorithmwaspresentedbyanItalyscholarM.Dorigolearningfromthebehaviorsimulatingantcolonyforaging.Firstly,thispaperhasintroducedthegroupintelligenceandpromotedtheantcolonyalgorithm,obtainedthechoiceregularof“m,α,β,ρ,Q”basingontheexperiment,andproposedaneffectivemethodnamed“foursteps”inthefundationofothersscholars’“threesteps”tochoosethemostsuperiorcombinationparameterofantgroupalgorithm,thenanaliedthesixkindsimprovedalgorithmofantcolonyantcolonyalgorithm,atthesametimeexplainedtheabilityofseveralkindsofantalgorithmtosolvetheTSPquestionaccordingtotheexperiments;obtainedthemostsuperiorresultaccordingtootherperformance(iterativenumberoftimes,iterativetimeandsoon)whenthemostsuperiorresultofTSPquestionoptimalsolutionissame.Finally,thispaperalsohascarriedontheoptimizationandtheimprovementtothevisibleprogrammingofChenYe’s“antcolonyalgorithmlaboratory”toenableitmoreconvenienttouseinseveralalgorithmsperformancecomparisonandthecomparisonofdifferentparameterandhomogeneousalgorithm.Keywords:Swarmintelligence;Antcolonyalgorithm;ParameterSelection;TSP;VisualizationIII目录1绪论............................................................................................................................11.1引言....................................................................................................................11.2群智能................................................................................................................11.3蚁群算法的提出................................................................................................21.4本文研究工作....................................................................................................22蚁群算法概述..........................................................................................................42.1蚁群算法基本原理.............................................................................................42.2蚁群算法的优点与不足.....................................................................................52.3本章小结............................................................................................................63蚁群算法的参数设置研究......................................................................................73.1硬件/软件环境平台...........................................................................................73.2蚂蚁数目对基本蚁群算法的影响....................................................................73.3信息启发式因子和期望值启发式因子............................................................93.4信息素残留系数............................................................................................103.5总信息量..........................................................................................................113.6本章小结..........................................................................................................124蚁群算法实验分析................................................................................................134.1改进的蚁群优化算法......................................................................................134.1.1最优解保留策略蚂蚁系统.......................................................................134.1.2蚁群系统...................................................................................................134.1.3最大-最小蚂蚁系统.................................................................................134.1.4基于排序的蚂蚁系统...............................................................................144.1.5TheBest-WorstAntSystem.....................................................................144.2实验仿真及算法性能比较分析.....................................................................154.2.1硬件/软件环境平台.................................................................................154.2.2重要参数设置...........................................................................................154.2.3实验结果...................................................................................................154.3本章小结..........................................................................................................175可视化编程............................................................................................................185.1“蚁群算法实验室”的优点与不足.....................................................................185.2最大最小蚁群算法的图形化演示的改进.......................................................185.3本章小结............................................................................................................226结论与展望........
本文标题: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...
链接地址:https://www.777doc.com/doc-30222 .html