多子种群微粒群免疫算法及其在函数优化中应用

《多子种群微粒群免疫算法及其在函数优化中应用》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、计算机研究与发展ISSN1000—1239／CN11-1777／TPJournalofComputerResearchandDevelopment49(9)：1883—1898，2012多子种群微粒群免疫算法及其在函数优化中应用吴建辉2章兢李仁发刘朝华(湖南大学电气与信息工程学院长沙410082)z(湖南大学信息科学与工程学院长沙410082)(wujianhui123@tom．com)AMulti—SubpopulationPSOImmuneAlgorithmandItsApplicationonFunctionOptimizationWuJianhui～，ZhangJing，LiRenfa

2、。，andLiuZhaohua(CollegeofElectricalandInformationEngineering，HunanUniversity，Changsha410082)(CollegeofInformationScienceandEngineering，HunanUniversity，Changsha410082)AbstractBasicparticleswarmoptimization(PSO)algorithm，whichisaglobalandparalleloptimizationofhighperformance，simplicity，robustness，nopr

3、oblemspecificinformation，etc．，hasbeenwidelyusedincomputerscience，optimizationofscheduling，functionoptimizationandotherfields．However，thebasicPSOalgorithmhasthedefectsofprematureconvergence，stagnationphenomenonandslowconvergencespeedinthelaterevolutionperiodforcomplexoptimizationproblems．Inordertoove

4、rcometheprematureconvergenceproblemofbasicPSOalgorithm，usingideaofmulti—subpopulationandself-adaptiveforreference，anovelmulti—subpopulationadaptivepolymorphiccrossbreedingparticleswarmoptimizationimmunealgorithm(MAPCPSOI)basedontwo—layermodelisproposed．Throughthebottomlayeradaptivepolymorphiccrossbr

5、eedingPSOoperationofseveralsubpopulations，theMAPCPSOIalgorithm，firstly，couldamelioratediversityofsubpopulationdistributionandeffectivelysuppressprematureandstagnationbehavioroftheconvergenceprocess．Secondly，theMAPCPSOIalgorithm，bythetoplayerimmuneclonalselectionoperationofseveralsubpopulations，could

6、significantlyimprovetheglobaloptimizationperformanceandfurtherenhancetheconvergenceprecision．ComparedwithotherimprovedPSOalgorithms，simulatedresultsoffunctionoptimizationshowthattheMAPCPSOIalgorithm，especiallysuitableforsolvinghigh—dimensionandmultimodaloptimizationproblems，hasrapiderconvergencespee

7、dandhighersolutionprecision．Keywordsparticleswarmoptimizationadaptive；polymorphicdiversityclonalselection；functionoptimization摘要为克服基本微粒群算法的早熟问题，借鉴多子种群和自适应的思想，提出了基于两层模型的多子种群自适应多态杂交微粒群免疫算法．该算法首先通过对若干个子种