Researchexperienceandacademicactivity，Whenasophomore,IjoinedtheAssociationofAIEnthusiastandbegantonarrowdownmyinterestformyfutureresearch.In1997,IparticipatedinsimulationtooldevelopmentfortheschedulingsysteminProf.Wang’slab.WiththetoolofOpenGLandMatlab,Idesignedasimulationprogramfortransportationschedulingsystem.ItisnowwidelyusedbydifferentresearchgroupsinNUST.In1998,Iassumedandfulfilledasewageanalysis&disposeprojectforNanjingsewagetreatmentplant.Thiswasmyfirstpracticetoconvertalaboratoryideatoacommercialproduct.

In1999,IjoinedthedistinguishedProfessorYu-GengXisresearchgroupaimingatNetworkflowproblemsolvingandHeuristicalgorithmresearch.SoonIwasengagedintheFuDanGeneDatabaseDesign.Mydutywastopickuptheusefulinformationamongdifferentkindsofgenematchingformat.Throughthecomparisonandanalysisformanyheuristicalgorithms,Iintroducedanimprovedevolutionaryalgorithm--Multi-populationGeneticAlgorithm.Byspanidingawholepopulationintoseveralsub-populations,thisimprovedalgorithmcaneffectivelypreventGAfromlocalconvergenceandpromotevariousevolutionaryorientations.ItprovedmoreefficientlythanSGAinexperiments,too.Inthesecondsemester,Ijoinedtheworkshop-schedulingresearchinShanghaiHeavyDutyTyreplant.Theschedulingwasdesignedfortherubber-makingprocessthatcoverednotonlydiscretebutalsocontinuouscircumstances.Tomakeabalancepointbetweenoptimizationqualityandtimecost,IproposedaDynamicLayeredSchedulingmethodbasedonhybridPetriNets.Thepracticalapplicationshowedthattheaveragemakespanwasshortenedbyalargescale.Ialsopublicizedtwopapersincorejournalswiththisidea.Recently,IamdoingresearchinthecompositePredictoftheElectricalPowersystemassistedwiththetechnologyofDataMiningforBaoSteel.ItrytocombinetheDecisionTreewithRecedingOptimizationtoprovideanewsolutionforthecompositePredictiveProblem.Thisprojectisnowunderconstruction.