简介:ONMECHANICALPROPERTYOFCONSTRAINTWeiYang(韦杨);LiangLifu(梁立孚);LiangZhongwei(梁忠伟)(ReceivedSep.6,1994;CommunicatedbyChienWeizang)A...
简介:
简介:Singularityanalysisinanimportantsubjectofthegeometricconstraintsatisfactionproblem.Inthispaper,threekindsofsingularitiesaredescribedandcorrespondingidentifcationmethodsarepresentedforbothunder0constrainedsystemsandover-constrainedsystems,Anotherspecialbutcommonsingularityforunder-constrainedgeometricsystems,pseudo-singularity,isanalyzed.Pseudo-singularityiscausedbyavarietyofconstraintmathchingofunder-constrainedsystemsandcanberemovedbyimprovingconstraintdistribution.Toavoidpseudo-singularityanddecideredundantconstraintsadaptively,adifferentiaitonalgorithmisproposedinthepaper.Itscorrctnessandeffciencyhavebeenvalidatedthroughitspracticalapplicationsina2D/3DgeometricconstraintsolverCBA.
简介:在这份报纸,我们学习一个乐队抑制nonnegative矩阵因式分解(乐队NMF)问题:为给定的nonnegative矩阵Y,与A作为Y斧子分解它一个nonnegative矩阵和X一个nonnegative块乐队矩阵。这个因式分解模型扩大一个单个低等级subspace模型到几重叠低等级subspaces的混合物,它能不仅提供稀少的表示,而且能从数据集捕获重要组织结构。基于重叠subspace聚类和在附近的subspaces之间的重叠的水平的俘获,二个简单、实际的算法被介绍解决乐队NMF问题。合成数据和真实图象数据的数字实验证明乐队NMF在数据表示提高NMF的性能并且处理。
简介:ByusingMoore-Penrosegeneralizedinverseandthegeneralsingularvaluedecompositionofmatrices,thispaperestablishesthenecessaryandsufficientconditionsfortheexistenceofandtheexpressionsforthecentrosymmetricsolutionswithasubmatrixconstraintofmatrixinverseproblemAX=B.Inaddition,inthesolutionsetofcorrespondingproblem,theexpressionoftheoptimalapproximationsolutiontoagivenmatrixisderived.
简介:在许多机器学习应用程序,数据不是免费的,并且为每个数据项目有测试费用。为节俭的原因,最小化测试的某存在工作尝试花费了并且同时,保存一个给定的决定系统的一个特别性质。在这份报纸,我们指出一个人能负担得起的测试费用在一些应用是有限的。因此,一个人不得不牺牲各自的性质在一项预算下面保留测试费用。形式化这期,我们定义测试费用限制属性减小问题,在优化目的是最小化有条件的信息熵的地方。这个问题是test-cost-sensitive属性减小问题和0-1背囊问题的必要归纳,因此它是更挑战性的。我们基于信息获得和测试费用建议一个启发式的算法处理新问题。算法在四UCI上被测试(加利福尼亚大学-Irvine)有各种各样的测试的数据集花费了背景。试验性的结果显示唯一的userspecified参数的适当设置。
简介:TheVaR,anewappearingfinancialrisk-managetool,havebeenappliedwidely.ManyfinancialsetupshaveaccustomedtomeasuretheriskofaportfoliowiththeVaR.SoitisverynecessarytodiscusstheportfoliochoiceproblemundertheVaRconstraint.Inthispaper,bysettingandsolvingtheportfoliochoicemodelundertheVaRconstraint,weillustratethattheuseoftheVaRconstraintreducesthearrayofchoicetoamoremanageablerange.TheprobabilityoftragetVaR,therefore,canbethoughtofasarisktoleranceassessmenttool(whencoupledwithanothermeasureofrisk).
简介:Motivatedbytheprojectsconstrainedbyspacecapacityandresourcetransportingtime,aprojectschedulingproblemwithcapacityconstraintwasmodeled.Ahybridalgorithmisproposed,whichusestheideasofbi-levelschedulingandprojectdecompositiontechnology,andthegeneticalgorithmandtabusearchiscombined.Topologicalreorderingtechnologyisusedtoimprovetheeffciencyofevaluation.Simulationresultsshowtheproposedalgorithmcanobtainsatisfiedschedulingresultsinacceptabletime.
简介:
简介:FortheARMAXsystemwithunknowncoefficientstheoptimaladaptivecontrolisdesignedsothatthefollowingrequirementsaremetsimultaneously:1)thetransferfunctionfromareferencesignaltothesystemoutputintheclosedloopequalsaprescribedrationalfunction;2)undertheconstraintmentionedin1)aquadraticlossfunctionisminimized;3)theparameterestimateisstronglyconsistent.
简介:
简介:明确的Bargmann对称限制被计算,它宽松的对的联系二进制nonlinearization为超级迪拉克系统被执行。在获得的对称限制下面,超级迪拉克层次的第n流动被分解成二个超级有限维的integrableHamiltonian系统,定义在上超级对称歧管有相应动态变量的R4N|2Nx和tn。运动的积分为Liouville要求integrability明确地被给。关键词对称限制-二进制nonlinearization-超级迪拉克系统-超级有限维的integrableHamiltonian系统2000苏布杰克特先生分类35Q51-37J35-37K10-37K40工程由Hangdian基础支持了(没有。KYS075608072),中国的国家自然科学基础(Nos.10671187,10971109)并且为中国的大学里的新世纪优秀才能的程序(没有。NCET-08-0515)。
简介:Inthetest-fieldcalibration,multi-azimuthstereoimagepairsareproducedoftheoutdoorlargecontrol-fieldbythestereo-visionsystemundercali-bration.Whileintheanalyticalprocessing,therelationshipbetweenimagepairsisadoptedasaconstraintcondition,whichensuresthestabilityandqualityofthecalibrationresults.Thispaperintroducesthedeductionprocessoftheconstraintconditions.
简介:Thispaperdealswiththeintegrabilityofafinite-dimensionalHamiltoniansystemlinkedwiththegeneralizedcoupledKdVhierarchy.ForthispurposetheassociatedLaxrepresentationispresentedafteranelementarycalculation.ItisshownthattheLaxrepresentationenjoysadynamicalr-matrixformulainsteadofaclassicaloneinthePoissonbracketonR2N.Consequentlytheresultingsystemisprovedtobecompletelyintegrableinviewofitsr-matrixstructure.
简介:TherearesomeadjustableparameterswhichdirecdyinfluencetheperformanceandstabilityofParticleSwarmOp-ttimizationalgorithm.Inthispaper,stabilitiesofPSOwithconstantparametersandtime-varyingparametersareanalyzedwithoutLipschitzconstraint.Necessaryandsufficientstabilityconditionsforaccelerationfactorψandinertiaweightwarepresented.Exper-imentsonbenchmarkfunctionsshowthegoodperfomanceofPSOsatisfyingthestabilitycondition,evenwithoutLipschitzcon-straint.Andtheinertiaweightwvalueisenhancedto(-1,1).