欢迎来到天天文库
浏览记录
ID:4163629
大小:651.11 KB
页数:5页
时间:2017-11-29
《一种蛙跳和差分进化混合算法》由会员上传分享,免费在线阅读,更多相关内容在行业资料-天天文库。
1、42011,47(18)ComputerEngineeringandApplications计算机工程与应用一种蛙跳和差分进化混合算法1,21,21何兵,车林仙,刘初升1,21,21HEBing,CHELinxian,LIUChusheng1.中国矿业大学机电工程学院,江苏徐州2210082.泸州职业技术学院机电工程研究所,四川泸州6460051.SchoolofMechanicalandElectricalEngineering,ChinaUniversityofMiningandTechnology,Xuzhou,Jiangsu221008,Ch
2、ina2.InstituteofMechatronicsEngineering,LuzhouVocationalandTechnicalCollege,Luzhou,Sichuan646005,ChinaHEBing,CHELinxian,LIUChusheng.Novelhybridshuffledfrogleapinganddifferentialevolutionalgorithm.Comput-erEngineeringandApplications,2011,47(18):4-8.Abstract:ShuffledLeapingFrogAl
3、gorithm(SFLA)ischaracterizedbysimplicity,fewcontrolparametersrequired,andeasilybeused,buthasthedisadvantagesofprematureconvergenceandlowprecisionforhardhigh-dimensionaloptimizationprob-lems,duetoitsrapidlossofthepopulationdiversityandthelackoflocalrefinedsearchabilitiesatthelat
4、erstagesofgen-erations.InordertoovercometheeasyprematureorearlyconvergenceofSFLAs,thispaperhybridizestheSFLAandtheDifferentialEvolution(DE)algorithmtoformahybridoptimizationalgorithm,namelySFL-DE,whichborrowstheideafromDE/best/1/binstrategythathastheadvantagesofstrongglobalsear
5、chabilityandbetterpopulationdiversity.Comparisonsarepresentedtotestperformancesofthenewalgorithmemploying6benchmark30-dimensionalfunctions.ComparedwithSFLAandstandardDE(i.e.,DE/best/1/binandDE/rand/1/binschemes)algorithms,theexperimentalresultsintermsoftheglobaloptimizationeffi
6、ciency,thesolutionaccuracyandthecomputationrobustnessdemonstratethattheSFL-DEalgorithmisabettertoolforsolvingsomebenchmarkoptimizationproblemswithinafewfixedgenerations,buttakesalongerruntime.Keywords:ShuffledFrogLeapingAlgorithm(SFLA);DifferentialEvolution(DE)algorithm;hybrido
7、ptimization;continuousoptimizationproblem摘要:混洗蛙跳算法(SFLA)具有算法简单、控制参数少、易于实现等优点,但在高维难优化问题中算法容易早熟收敛且求解精度不高。导致该缺陷的主要原因是在进化后期种群多样性迅速下降,且缺乏局部细化搜索能力。借鉴差分进化算法(DE)中DE/best/1/bin版本具有全局搜索能力较强、种群多样性较好的优点,将SFLA与DE有机融合,形成混合优化算法(SFL-DE),以克服SFLA容易早熟收敛的缺陷。给出了6个30维benchmark问题数值对比实验,结果表明,在给定的较小进化代
8、数内,SFL-DE的寻优效率、计算精度、鲁棒性等性能优于SFLA和基本DE(DE/best/1/bin和DE
此文档下载收益归作者所有