简介：Inthispaper,wepresentanovelSupportVectorMachineactivelearningalgorithmforeffective3Dmodelretrievalusingtheconceptofrelevancefeedback.Theproposedmethodlearnsfromthemostinformativeobjectswhicharemarkedbytheuser,andthencreatesaboundaryseparatingtherelevantmodelsfromirrelevantones.Whatitneedsisonlyasmallnumberof3Dmodelslabelledbytheuser.Itcangrasptheuser'ssemanticknowledgerapidlyandaccurately.Experimentalresultsshowedthattheproposedalgorithmsignificantlyimprovestheretrievaleffectiveness.Comparedwithfourstate-of-the-artqueryrefinementschemesfor3Dmodelretrieval,itprovidessuperiorretrievalperformanceafternomorethantworoundsofrelevance

简介：TheTerabyteAnalysisMachineProjectisDevelopinghardwareandsoftwaretoanalyzeTerabytescaledatasets.TheDistanceMachineframeworkprovidesfacilitiestoflexiblyinterfaceapplicationspecificindexingandpartitioningalgorthmstolargescientificdatabases.

简介：Ininjectionmouldingproduction,thetuningoftheprocessparametersisachallengingjob,whichreliesheavilyontheexperienceofskilledoperators.Inthispaper,takingintoconsiderationoperatorassessmentduringmouldingtrials,anovelintelligentmodelforautomatedtuningofprocessparametersisproposed.Thisconsistsofcasebasedreasoning(CBR),empiricalmodel(EM),andfuzzylogic(FL)methods.CBRandEMareusedtoimitaterecallandintuitivethoughtsofskilledoperators,respectively,whileFLisadoptedtosimulatetheskilledoperatoroptimizationthoughts.First,CBRisusedtosetuptheinitialprocessparameters.IfCBRfails,EMisemployedtocalculatetheinitialparameters.Next,amouldingtrialisperformedusingtheinitialparameters.ThenFLisadoptedtooptimizetheseparametersandcorrectdefectsrepeatedlyuntilthemouldedpartisfoundtobesatisfactory.Basedontheabovemethodologies,intelligentsoftwarewasdevelopedandembeddedinthecontrollerofaninjectionmouldingmachine.Experimentalresultsshowthattheintelligentsoftwarecanbeeffectivelyusedinpracticalproduction,anditgreatlyreducesthedependenceontheexperienceoftheoperators.

简介：在这篇论文，我们以速度在二台一致机器上调查安排问题的变体1；s。为这个问题，我们被给二台潜在的一致机器处理独立工作的一个序列。机器需要在开始到过程前被激活，；激活的每台机器招致固定机器激活费用。没有机器开始被激活，；当一个工作被揭示时，算法有选择激活新机器。目的是最小化makes平底锅的和；机器激活费用。我们与竞争比率设计最佳的联机算法(2s+1)为每s≥1的/(s+1)。

简介：

简介：Thisworkdealswiththecoolingofhigh-speedelectricmachines,suchasmotorsandgenerators,throughanairgap.Itconsistsofnumericalandexperimentalmodelingofgasflowandheattransferinanannularchannel.Velocityandtemperatureprofilesaremodeledintheairgapofahigh-speedtestmachine.Frictionandheattransfercoefficientsarepresentedinalargevelocityrange.Thegoalsarereachedacceptablyusingnumericalandexperimentalresearch.Thevelocityfieldbythenumericalmethoddoesnotmatchineveryrespecttheestimatedflowmode.TheabsenceofsecondaryTaylorvorticesisevidentwhenusingtimeaveragednumericalsimulation.

简介：LECR4isthefirstECRionsourceusingevaporativecoolingtechnologyintheworld.Itsuniquefeatureisthatthesolenoidsweremadefromsolidsquarecopperwires(3.32mm×5.77mmwithinsulation),andthatallthecoilsareentirelyimmersedintheroomtemperaturecoolant.Thecoilscanproduceamaximumaxialmagneticfieldupto2.5T.LECR4aimstoprovideintensemultiplechargestateionbeamsforSSC-Linacproject-anewLinearinjectorfortheSeparatedSectorCyclotron(SSC).Presently,someionbeamswithdifferentM/QratiohavebeenacceleratedsuccessfullywiththeRFQandDTL,suchas200eμAof16O^5+,200eμAof40Ar^8+,50eμAof209Bi^30+etc.ThemeasuredtransmissionefficiencyofRFQisupto90%.ThelayoutofLECR4ionsourceandtheLEBTisshowninFig.1.

简介：Thispaperpresentsanapplicationofiterativelearningcontrol(ILC)techniquetothevoltagecontrolofsolidoxidefuelcell(SOFC)stack.Tomeetthedemandsofthecontrolsystemdesign,anautoregressivemodelwithexogenousinput(ARX)isestablished.Firstly,byregulatingthevariationofthehydrogenflowrateproportionaltothatofthecurrent,thefuelutilizationoftheSOFCiskeptwithinitsadmissiblerange.Then,basedontheARXmodel,threekindsofILCcontrollers,i.e.P-,PI-andPD-typearedesignedtokeepthevoltageatadesiredlevel.SimulationresultsdemonstratethepotentialoftheARXmodelappliedtothecontroloftheSOFC,andprovetheexcellenceoftheILCcontrollersforthevoltagecontroloftheSOFC.

简介：Theprocessingmethodappliedtothesidesurfaceisdifferentfromthemethodappliedtothelightpasssurfaceinneodymiumphosphateglass(Nd:glass),andthussubsurfacedefectsremainafterprocessing.ThesubsurfacedefectsinthesidesurfaceinfluencethegainuniformityofNd:glass,whichisakeyfactortoevaluatetheperformanceofamplifiers.Thescatteringcharacteristicsofsidesubsurfacedefectsweresimulatedbyfinitedifferencetimedomain(FDTD)Solutionssoftware.Thescatteringpowersoftheglassfabricatedbyacomputernumericalcontrol(CNC)machinewithoutcladdingweretestedatdifferentincidentangles.Thetrendofthecurvewassimilartothesimulatedresult,whilethesmallestpointwasdifferentwiththecomplextruemorphology.Thesimulationshowedthattheequivalentresidualreflectivityofthecladdingglasscanbemorethan0.1%whenthenumberofdefectsinasinglegriddingisgreaterthan50.

简介：

简介：<正>Someproblemsinusingv-supportvectormachine(v-SVM)forthepredictionofnonlineartimeseriesarediscussed.Theproblemsincludeselectionofvariousnetparameters,whichaffecttheperformanceofprediction,mixtureofkernels,anddecompositioncooperationlinearprogrammingv-SVMregression,whichresultinimprovementsofthealgorithm.ComputersimulationsinthepredictionofnonlineartimeseriesproducedbyMackey-GlassequationandLorenzequationprovidesomeimprovedresults.

简介：Thispaperproposesalearning-basedmethodfortextdetectionandtextsegmentationinnaturalsceneimages.First,theinputimageisdecomposedintomultipleconnected-components(CCs)byNiblackclusteringalgorithm.ThenalltheCCsincludingtextCCsandnon-textCCsareverifiedontheirtextfeaturesbya2-stageclassificationmodule,wheremostnon-textCCsarediscardedbyanattentionalcascadeclassifierandremainingCCsarefurtherverifiedbyanSVM.AlltheacceptedCCsareoutputtoresultintextonlybinaryimage.Experimentswithmanyimagesindifferentscenesshowedsatisfactoryperformanceofourproposedmethod.

简介：TheZ-backlighterlaserfacilityprimarilyconsistsoftwohighenergy,high-powerlasersystems.Z-Beamletlaser(ZBL)(Ramboetal.,Appl.Opt.44,2421(2005))isamulti-kJ-class,nanosecondlaseroperatingat1054nmwhichisfrequencydoubledto527nminordertoprovidex-raybacklightingofhighenergydensityeventsontheZ-machine.Z-Petawatt(ZPW)(Schwarzetal.,J.Phys.:Conf.Ser.112,032020(2008))isapetawatt-classsystemoperatingat1054nmdeliveringupto500Jin500fsforbacklightingandvariousshort-pulselaserexperiments(seealsoFigure10forafacilityoverview).Withthedevelopmentofthemagnetizedlinerinertialfusion(MagLIF)conceptontheZ-machine,theprimarybacklightingmissionsofZBLandZPWhavebeenadjustedaccordingly.Asaresult,wehavefocusedourrecenteffortsonincreasingtheoutputenergyofZBLfrom2to4kJat527nmbymodifyingthefiberfrontendtonowincludeextrabandwidth(forstimulatedBrillouinscatteringsuppression).TheMagLIFconceptrequiresawell-defined/behavedbeamforinteractionwiththepressurizedfuel.HencewehavemadegreateffortstoimplementanadaptiveopticssystemonZBLandhaveexploredtheuseofphaseplates.WearealsoexploringconceptstouseZPWasabacklighterforZBLdrivenMagLIFexperiments.Alternatively,ZPWcouldbeusedasanadditionalfusionfuelpre-heaterorasatemporallyflexiblehighenergypre-pulse.AlloftheseconceptsrequiretheabilitytooperatetheZPWinananosecondlong-pulsemode,inwhichthebeamcanco-propagatewithZBL.Someoftheproposedmodificationsarecompleteandmostofthemarewellontheirway.

简介：