简介：作者与记忆显示出随机的波浪方程的一个班的答案的存在和唯一。解决方案的精力函数的腐烂估计也被获得。有存储器的关键词随机的波浪方程-溶剂-无限的维的维纳过程-精力功能2000苏布杰克特先生分类60H15-35L05工程由国家天赋支持了中国(号码10871103)的科学基础。

简介：Theauthorsstudyrecursivestructureswith'finitememory'inEuclideanmatricspaceandthemultifractaldecompositionofthecorrespondingfractals.Foranytwopositivenumbersq,β,andsucharecursivestructure,alinearoperatorV~(q,β)infinitedimensionalspaceisdefined.ThemultifractalspectrumisgivenbythespectralradiusofV~(q,β).

简介：Inthispaper,westudytheasymptoticCUSUMtestsfordetectingchangesinthemeanorvarianceofamoving-averageprocesswithlongmemory.Whenthereisnochangeover[O,T],theasymptoticdistributionoftheteststatisticisderived,whichallowsustofindasymptoticcriticalvalues.Whenthereisachange,thebehavioroftheteststatisticisdiscussed.Conditionsfortheconsistencyofthesetestsarealsodiscussed.Basedontheasymptoticresults,simulationstudiesoftestingforchangesinthemeanshowthattheCUSUMtestproposedperformswell.

简介：<正>Supposethat{αk}k∞=-∞isaLacunarysequenceofpositivenumberssatisfyinginfαk+1/αk=α>1andthatΩ（y’）isafunctionintheBesovspaceB10,1(Sn-1)whereSn-1istheunitsphereonRn（n≥2）.Weprovethatif∫Sn-1Ω（y’）dσ（y’）=0thenthediscretesingularintegraloperatorTΩf（x）=∫(Sn-1)f（x-αky’）Ω（y’）dσ（y’）andtheassociatedmaximaloperatorTΩf（x）=∫Sn-1f（x-αky’）Ω（y’）dσ（y’）arebothboundedinthespaceL2（Rn）.ThetheoremsinthispaperimprovearesultbyDuoandikoetxeaandRubiodeFranciaintheL2case.

简介：Theauthorsgivetheexistenceresultstothenonlineardifferentialinclusionu'(t)∈Au(t)+F(t,ut),whereAisagangratorofequicontinuoussemigroupandFismultivalued.

简介：WestudyhowtousetheSR1updatetorealizeminimizationmethodsforproblemswherethestorageiscritical.Wegiveanupdateformulawhichgeneratesmatricesusinginformationfromthelastmiterations.Thenumericaltestsshowthatthemethodisefficent.

简介：LetMfbetheKakeyamaximalfunctionind-dimensionalEuclideanspace,withsamebaseconsistingofcylindersofeccentricityN.Theinequalityshoumforabasesatisfyingadirectioncondition,where?andcareconstantsdependingonlyond.

简介：支持向量机器原来为二进制分类被设计。怎么有效地扩大它因为多班分类仍然是一个进行中的研究问题。Inthis纸，我们认为是多范畴支持的自然延期的核机器是Crammer和歌手原来建议的向量机器。基于算法稳定性，我们为机器在论文建议了的核获得归纳错误界限。

简介：在这篇论文，我们在不同计算设定在n宽度上从空格l_p(1≤p≤2)为斜操作符T把一些最佳的算法给l_2。