hthecomparisonandanalysisformanyheuristicalgorithms,iintroducedanimprovedevolutionaryalgorithm--multi-populationgeneticalgorithm.bydividingawholepopulationintoseveralsub-populatio,thisimprovedalgorithmcaneffectivelypreventgafromlocalconvergenceandpromotevariousevolutionaryorientatio.itprovedmoreefficientlythansgainexperiments,too.inthesecondsemester,ijoinedtheworkshop-schedulingresearchinshanghaiheavydutytyreplant.theschedulingwasdesignedfortheruer-makingprocethatcoverednotonlydiscretebutalsocontinuouscircumstances.tomakeabalancepointbetweenoptimizationqualityandtimecost,iproposedadynamiclayeredschedulingmethodbasedonhybridpetrinets.thepracticalalicationshowedthattheaveragemakeanwasshortenedbyalargescale.ialsopublicizedtwopapersincorejournalswiththisidea.recently,i'amdoingresearchinthecompositepredictoftheelectricalpowersystemaistedwiththetechnologyofdataminingforbaosteel.itrytocombinethedecisiontreewithrecedingoptimizationtoprovideanewsolutionforthecompositepredictiveproblem.thisprojectisnowundercotruction.
　　besides,injuly2000,igottheoortunitytogivealectureinenglishinasiacontrolconference(ascc)whichisoneofthetop-levelconferencesamongtheworldintheareaofcontrolandautomation.inmysenioryear,imetprof.xiao-songlin,avisitingprofeorofmathematicsfromuniversityofcalifornia-riverside,ilearnedgraphtheoryfromhimformynetworkresearch.theseexperiencesallrapidlyexpandedmyknowledgeofenglishandtheunderstandingofwesternculture.
　　ihopetostudyindepth
　　inretroect,ifindmyselfstandingonasolidbasisinboththeoryandexperience,whichhaspreparedmefortheph.d.program.myfutureresearchinterestsinclude:networkschedulingproblem,heuristicalgorithmresearch(eeciallyingaandneuralnetwork),sulychainnetworkresearch,hybridsystemperformanceanalysiswithpetrinetsanddatamining.
　　pleasegivemyalicationmaterialsaseriouscoideration.thankyouverymuch.
　　

上一页  [1] [2]