粗糙集并行约简算法研究

《粗糙集并行约简算法研究》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、RESEARCHONALGORITHMSOFPARALLELREDUCTSnqROUGHSETSABSTRACTBydefiningtheconceptsofupperapproximationandlowerapproximationroughsetsmakethevaguedegreecomputableSoithasparticularadvantageswhenitwasusedtodealwithincompleteandimprecisedata：1)Needingnoprioriknowledge；2)Moreobjecti

2、veandprecise；31Highlycomplementary'withothertheoriesofuncer【aintyDuetotheseadvantagesroughsetstheoryhasbeenfullydevelopedandwideb’usedsinceitwasproposedabout30yearsagoDuringmanyyearsdevelopmenttheavailabilityofroughsettheoryhasbeenfully、erifiedandtheoryframeworkhasbeenimp

3、rovedaswellButthechallenge．howtoobtainastablereductfromadynamicorincreasingdataSethasnotbeensuccessfullysolvedAlthoughdynamicreducttheoryhaspointedolltaverygoodwaytosolvetheissuebutduetoitstimecomplexityofexponentialorder．dynamicreductscouldonlybeusedtoobtainreductsfromth

4、eless-conditionalattributedataTheapplicationofdynamicreductswillbelimitedMoreoverthemethodofdynamicreductsisthegreatlyincompletebecauseintersectionofallofreductsofafamilydecisionsubsystemsmaybeerupt3Tosolvetheissuethathowtoobtainastablereduetfromadynamicorincreasingdatase

5、tDayongDengproposedtheparallelreducttheorySimilartodynamicreductsparallelreductsdivideadecisionsystemintoseveraldecisionsubsystemsandreduceredundantconditionanribute(s)fromthemsimuhaneousl，vfinallyoneormorestablereduct(s)wouldbeobtainedfromthefamilyofdecisionsubsystemsDif

6、ferentfromdynamicreductalgorithm。sexhaustivesearchingamongdecisionsubsystems，parallelreductscanJudgewhetheraconditionattributeismoreimpoaantthanotherswithheuristicinformationintheprocessingofobtainingparallelreductssothetimecompiexit)

7、ofparagelreductalgorithmispolynomialT

8、heparallelreducttheor)isacompletetheoDbecauseitcanavoidtheemptyresultfromitsdefinitionBesides，parallelreducttheor3Callmakeuseofallachievementsofroughsetsbutdynamicreducttheor)cantl】IInthisthesisWeinvestigateparallelreducttheorydeeply,andproposeanimprovedaparallelreduclalg

9、orithmThe㈣algorithmcanobtainaparallelreductwhichcontainsfewereondgionanributeswithinthesametimec