简介:<正>Customerdataisrevolution-izingasaglobaltrendandChinaneedstofollowthetrend,saidabooknamed"InDataWeTrust"translatedbyRolandBergerStrategyConsultantsandChinaMachinePress."Thecustomerdatarevolutionhasmovedtothenextlevel.Tenyearsago,onlybigcompaniescouldaffordtoen-gageindatamining;today,everypizzadeliverycompanycandoit.Thedaysof
简介:Thispaperpresentsasimplenonparametricregressionapproachtodata-drivencomputinginelasticity.Weapplythekernelregressiontothematerialdataset,andformulateasystemofnonlinearequationssolvedtoobtainastaticequilibriumstateofanelasticstructure.Preliminarynumericalexperimentsillustratethat,comparedwithexistingmethods,theproposedmethodfindsareasonablesolutionevenifdatapointsdistributecoarselyinagivenmaterialdataset.
简介:有效职员rostering并且遇见门诊病人需求的耐心的安排是一项很复杂、动态的任务。由于受欢迎的变化和专家可获得性,专家分配一定很灵活、非近视。医药专家典型地在亚专门化被限制,服务几个病人组织并且是在对诊所和操作空间的耐心的访问的链的关键资源(或)。克服一个近视看法一次离开安排的约会,当分配关键资源到不同的耐心的组时,我们通过耐心的约会的链探讨耐心的流动。我们在场到花名同活动和病人的专家的自动分配的一条新、数据驱动的算法的途径组织。由他们的很自然,简化数学模型不能捕获对被建模的系统特征的复杂性。在我们的途径,到他们的日常活动的专家的分配对过去和现在的关键资源可获得性,以及到过去的资源分配灵活、应答。在花名同活动的可变性活跃地被最小化,以便提高供应链流动。与用实验数据的申请盒子的分离事件的模拟,我们说明我们的途径怎么改进耐心的服务水平(SL,病人的百分比服务了在时间上)象一样等待时间(天),没有在资源能力的变化。
简介:Largesizedpowertransformersareimportantpartsofthepowersupplychain.Theseverycriticalnetworksofengineeringassetsareanessentialbaseofanation'senergyresourceinfrastructure.Thisresearchidentifiesthekeyfactorsinfluencingtransformernormaloperatingconditionsandpredictstheassetmanagementlifespan.Engineeringassetresearchhasdevelopedfewlifespanforecastingmethodscombiningreal-timemonitoringsolutionsfortransformermaintenanceandreplacement.Utilizingtherichdatasourcefromaremoteterminalunit(RTU)systemforsensor-datadrivenanalysis,thisresearchdevelopsaninnovativereal-timelifespanforecastingapproachapplyinglogisticregressionbasedontheWeibulldistribution.Themethodologyandtheimplementationprototypeareverifiedusingadataseriesfrom161kVtransformerstoevaluatetheefficiencyandaccuracyforenergysectorapplications.Theassetstakeholdersandsupplierssignificantlybenefitfromthereal-timepowertransformerlifespanevaluationformaintenanceandreplacementdecisionsupport.
简介:VelocityInterferometerSystemforAnyReflector(VISAR)[BarkerandHollenbach,J.Appl.Phys.43,4669(1972)]isawell-knowndiagnosticthatisemployedonmanyshockphysicsandpulsed-powerexperiments.WiththeVISARdiagnostic,thevelocityonthesurfaceofanymetalflyercanbefound.FormostexperimentsemployingVISAR,eitherakineticpressure[Grady,Mech.Mater.29,181(1998)]oramagneticpressure[Lemkeetal.,IntlJ.ImpactEng.38,480(2011)]drivesthemotionoftheflyer.Moreover,reliablepredictionofthetime-dependentpressureisoftenacriticalcomponenttounderstandingthephysicsoftheseexperiments.AlthoughVISARcanprovideaprecisemeasurementofaflyer’ssurfacevelocity,therealchallengeofthisdiagnosticimplementationisusingthisvelocitytounfoldthetimedependentpressure.ThepurposeofthispaperistoelucidateanewmethodforquicklyandreliablyunfoldingVISARdata.
简介:Anopticalfibertransmittingphotoelectriccouplingcurrenttransformerhasbeendeveloped.Intheenvironmentofhighvoltageandelectromagneticfields,thisdeviceischaracterizedbyhighinsulation,highreliability,smallmassandhighanti-interference.Thestructuredesignandopertingprincipleofthesystemareintroduced,measurementerrorsareanalyzedandexperimentalresultsaregiven.
简介:Wedevelopadatadrivenmethod(probabilitymodel)toconstructacompositeshapedescriptorbycombiningapairofscale-basedshapedescriptors.Theselectionofapairofscale-basedshapedescriptorsismodeledasthecomputationoftheunionoftwoevents,i.e.,retrievingsimilarshapesbyusingasinglescale-basedshapedescriptor.Thepairofscale-basedshapedescriptorswiththehighestprobabilityformsthecompositeshapedescriptor.Givenashapedatabase,thecompositeshapedescriptorsfortheshapesconstituteaplanarpointset.AVoR-Treeoftheplanarpointsetisthenusedasanindexingstructureforefficientqueryoperation.Experimentsandcomparisonsshowtheeffectivenessandefficiencyoftheproposedcompositeshapedescriptor.
简介:
简介:Thispaperdealswithanimportantsubjectofrejectingtheabnormaldataintelligentlyinthedynamicdatasystem.Basedontheprincipleofnearestneighboroffuz-zymathematics,anapproachofmathematicallyabstractingthehumanthinkingandphys-icalpracticeknowledgeisdiscussed,anewmethodofautomaticrejectionofabnormalda-taisthenproposed.Theexperimentalresultsshowthatthemethodisavailabletothepractice.
简介:ThispaperpresentsanotherformalproofforthecorrectnessoftheDeadlineDrivenScheduler(DDS).ThisproofisgivenintermsofDurationCalculuswhichprovidesabstractionforrandompreemptionofprocessor.Comparedwithotherapproaches,thisproofreliesonmanyintuitivefacts.Thereforethisproofismoreintuitive,whileitisstillformal.
简介:ThepaperestablishestwostochasticSIRSmodelswithjumpstodescribethespreadofnetworkvirusbycyberwar,terrorismandothers.First,addingrandomperturbationsproportionallytoeachvariable,wegetthedynamicpropertiesaroundthepositiveequilibriumofthedeterministicmodelandtheconditionsforpersistenceandextinction.Second,givingarandomdisturbancetoendemicequilibrium,wegetastochasticsystemwithjumps.BymodifyingtheexistingLyapunovfunction,weprovethepositivesolutionofthesystemisstochasticallystable.
简介:CoordinatesofthekeyfacialfeaturepointscanbecapturedbymotioncapturesystemOPTOTRAKwithreal-timecharacterandhighaccuracy.Thefacialmodelisconsideredasanundirectedweightedgraph.Byiterativelysubdividingtherelatedtriangleedges,thegeodesicdistancebetweenpointsonthemodelsurfaceisfinallyobtained.TheRBF(RadialBasisFunctions)interpolationtechniquebasedongeodesicdistanceisappliedtogeneratedeformationofthefacialmeshmodel.Experimentalresultsdemonstratethatthegeodesicdistancecanexplorethecomplextopologyofhumanfacemodelsperfectlyandthemethodcangeneraterealisticfacialexpressions.
简介:Inthispaper,anequationsystemofelectrohydrodynamics(EHD)basedonthefluidmotionequationsisdiscussed.Emphasisisputontheeffectsofelectricalforceandsurfacetensionuponthefluidmotion.TheMarkerandCellmethodisusedtosetupacomputationalsimulationprogramofelectricallydrivenmotionoffluid.Withthehelpoftheprogram,acylindricalfluidundertheinfluencesofelectricalfieldandsurfacetensionhasbeencalculated.Theresultisingoodagreementwiththeexperimentalobservation.