基于量子精英蛙的最小属性自适应合作型协同约简算法.pdf

《基于量子精英蛙的最小属性自适应合作型协同约简算法.pdf》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DOI：10．7544／issnlOOO一1239．2014．20120572JournalofComputerResearchandDevelopment51(4)：743—753，2014基于量子精英蛙的最小属性自适应合作型协同约简算法丁卫平。王建东管致锦(南通大学计算机科学与技术学院江苏南通226019)(南京航空航天大学计算机科学与技术学院南京210016)。(计算机软件新技术国家重点实验室(南京大学)南京210093)(ding．wp@ntu．edu．cn)AMinimumAttributeSelf__AdaptiveCooperat

2、iveCo--EvolutionaryReductionAlgorithmBasedonQuantumElitistFrogsDingWeiping’～，WangJiandong。，andGuanZhijin(SchoolofComputerScienceandTechnology，NantongUniversity，Nantong，Jiangsu226019)。(College0／ComputerScienceandTechnology，NanjingUniversityofAeronauticsandAstronautics，Nanjing210016)。

3、(StateKeyLaboratoryforNovelSc)ftwareTechnology(NanjingUniversity)．Nanjing210093)AbstractAttributereductionisakeypointinstudyingroughsetstheory．IthasbeenproventhatcomputingminimumattributereductionofthedecisiontableiSanNP—hardproblem．However。theconventionalevolutionaryalgorithmsareno

4、tefficientinaccomplishingminimumattributereduction．Anovelminimumattributeself—adaptivecooperativeCO—evolutionaryreductionalgorithm(QEFASCR)basedonquantumelitistfrogsisproposed．Firstly，evolutionaryfrogsarerepresentedbymulti—statequantumchromosomes，andquantumelitistfrogscanfastguideth

5、eevolutionaryfrogsintotheoptimalarea，whichcanstrengthentheconvergencevelocityandglobalsearchefficiency．Secondly，aself-adaptivecooperativeCO—evolutionarymodelforminimumattributereductionisdesignedtodecomposeevolutionaryattributesetsintoreasonablesubsetsaccordingtoboththebesthistorica

6、1performanceexperiencerecordsandassignmentcredits，andsomeoptimalelitistsindifferentsubpopuIationsareselectedouttoevolvetheirrespectiveattributesubsets，whichcanincreasethecooperationandefficiencyofattributereduction．Thereforetheglobalminimumattributereductionsetcanbeobtainedquickly．E

7、xperimentsresultsindicatethattheproposedalgorithmcanachievethehigherperformanceontheefficiencyandaccuracyofminimumattributereduction，comparedwiththeexistingalgorithms．Keywordsminimumattributereduction；quantumelitistfrog；cooperativeCO～evolution；self-adaptivedecomposition；bestperforma

8、nceexperience；assig