简介:AbstractAcyclichypergraphsareanaloguesofforestsingraphs.Theyareveryusefulinthedesignofdatabases.Inthisarticle,themaximumsizeofanacvclichypergraphisdeterminedandthenumberofmaximumγ-uniformacyclichypergraphsofordernisshowntobe(r-1n)(n(r-1)-r2+2r)n-r-1.
简介:Inordertoconstructestimatingfunctionsinsomeparametricmodels,thispaperintroducestwoclassesofinformationmatrices.Somenecessaryandsufficientconditionsfortheinformationmatricesachievingtheirupperboundsaregiven.Fortheproblemofestimatingthemedian,someoptimumestimatingfunctionsbasedontheinformationmatricesareacquired.Undersomeregularityconditions,anapproachtocarryingoutthebestbasisfunctionisintroduced.Innonlinearregressionmodels,anoptimumestimatingfunctionbasedontheinformationmatricesisobtained.Someexamplesaregiventoillustratetheresults.Finally,theconceptofoptimumestimatingfunctionandthemethodsofconstructingoptimumestimatingfunctionaredevelopedinmoregeneralstatisticalmodels.
简介:Anewmethodoftreatingmaximumwaveheightasarandomvariableinreliabilityanalysisofbreakwatercaissonsisproposed.Themaximumwaveheightisexpressedasthesignificantwaveheightmultipliedbytheso-calledwaveheightratio.Theproposedwaveheightratioisatypeoftransferfunctionfromthesignificantwaveheighttothemaximumwaveheight.Undertheconditionofabreakingwave,theratioisintrinsicallynonlinear.Therefore,theprobabilitydensityfunctionforthe
简介:LetP(z)beapolynomialofdegreenhavingallitszerosin|z|≤k,k≤1,thenforeveryrealorcomplexnumberβ,with|β|≤1andR≤1,itwasshownbyA.Zirehetal.[7]thatfor|z|=1,min|z|=1|P(Rz)+β(R+k/1+k)^nP(z)|≥k^-n|Rn+b(R+k/1+k)^n|min|z|=k|P(z)|.Inthispaper,weshallpresentarefinementoftheaboveinequality.Besides,weshallalsogeneralizesomewell-knownresults.
简介:Amaximumprincipleisprovedforsemilinearstochasticevolutionsystems.Themaincontributionofthisworkisthatinourproblem,theinfinitesimalgeneratorofthesemigroupofthesystemsneednottobeelliptic.ThisgeneralizesaresultofA.Bensoussanin1983.
简介:Inthispapertheauthorsgeneralizetheclassicrandombipartitegraphmodel,anddefineamodeloftherandombipartitemultigraphsasfollows:letm=m(n)beapositiveinteger-valuedfunctiononnandζ(n,m;{pk})theprobabilityspaceconsistingofallthelabeledbipartitemultigraphswithtwovertexsetsA={a_1,a_2,...,a_n}andB={b_1,b_2,...,b_m},inwhichthenumberst_(ai),b_joftheedgesbetweenanytwoverticesa_i∈Aandb_j∈BareidenticallydistributedindependentrandomvariableswithdistributionP{t_(ai),b_j=k}=pk,k=0,1,2,...,wherepk≥0and∞Σk=0pk=1.TheyobtainthatX_(c,d,A),thenumberofverticesinAwithdegreebetweencanddofG_(n,m)∈ζ(n,m;{pk})hasasymptoticallyPoissondistribution,andanswerthefollowingtwoquestionsaboutthespaceζ(n,m;{pk})with{pk}havinggeometricdistribution,binomialdistributionandPoissondistribution,respectively.Underwhichconditionfor{pk}cantherebeafunctionD(n)suchthatalmosteveryrandommultigraphG_(n,m)∈ζ(n,m;{pk})hasmaximumdegreeD(n)inA?underwhichconditionfor{pk}hasalmosteverymultigraphG(n,m)∈ζ(n,m;{pk})auniquevertexofmaximumdegreeinA?
简介:Parameterizationisoneofthekeyproblemsintheconstructionofacurvetointerpolateasetoforderedpoints.Weproposeanewlocalparameterizationmethodbasedonthecurvaturemodelinthispaper.Thenewmethoddeterminestheknotsbyminimizingthemaximumcurvatureofquadraticcurve.Whentheknotsbythenewmethodareusedtoconstructinterpolationcurve,theconstructedcurvehavegoodprecision.Wealsogivesomecomparisonsofthenewmethodwithexistingmethods,andourmethodcanperformbetterininterpolationerror,andtheinterpolatedcurveismorefairing.
简介:在这研究,越过从1956~2004的中国的最大的风速度(WSmax)变化基于观察车站数据被分析,并且为20462065和20802099的WSmax的变化用三个全球气候模型被投射(GFDL_CM2_0,CCCMA_CGCM3,并且MRI_CGCM2)那参予了IPCC第四份评价报告(AR4)。观察到年度、季节的WSmax和强风天的频率显示出明显的衰退趋势。年度WSmax每十年由约1.46ms1减少了,并且强风天的数字从1956~2004每十年到3.0天减少了。年度、季节的WSmax减少的振幅比年度、季节的平均的风速度(WSavg)的那些大。变弱东方亚洲冬季和夏天季风在整个中国上是为WSmax和WSavg的不同减少的原因。在东南的WSmax的减少中国的沿海的区域与在中国和减少的数字的冷波浪的减少的紧张有关(并且减少的紧张)land-falling,台风在西北太平洋发源。全球气候模型GFDL_CM2_0,MRI_CGCM2,和EBGCM(上面的整体提及三个全球气候模型)一致地建议年度、季节的WSmax价值将相对19812000在20462065和20802099期间减少。模型也建议在为整个中国的WSmax的减少在20462065和20802099期间与冷波浪的减少的紧张和冬季季风的减少的紧张有关,并且在在东南的WSmax的减少中国的沿海的区域在一样的时期期间相应于在在夏天的西北太平洋上的热带气旋的减少的数字。
简介:Themaximummatchinggraphofagraphhasavertexforeachmaximummatchingandanedgeforeachpairofmaximummatchingswhichdifferbyexactlyoneedge.Inthispaper,weobtainalowerboundofdistancebetweentwoverticesofmaximummatchinggraph,andgiveanecessaryandsufficientconditionthattheboundcanbereached.
简介:让G没有固定的点在的一个集合上的一个排列组,和m是一个积极整数。然后,G的运动被定义为行动(G):=supG{|Gg\G|g?G}行动(G):=\mathop{\sup}\limits_\Gamma\left\{{\left|{\Gamma^g\backslash\Gamma}\right|\left|{g\inG}\right}\right\}。它被Praeger显示出如果行动(G)=m那时|3m+t1,在t是G轨道在上的数字的地方。在这份报纸,有有最大值的度3m+t1的所有不及物的排列组跳了被分类。确实,她问题的一个积极答案是否上面的界限|=3m+t1为|为每t是锋利的>1被给。
简介:线性假设是最大的可能性的主要劣势线性re-gressionMLLR。这份报纸使用多项式回归方法为改编建模并且为柔韧的语音识别用最大的可能性的多项式回归MLPR建立一个非线性的模型改编算法。在这个算法,在在每条Mel隧道训练并且测试Gaussian工具之间的非线性的关系被一套多项式接近,多项式系数用期望在测试envi-ronment从改编数据被估计--最大化他们算法和最大的可能性的ML标准。试验性的结果证明秒顺序多项式能接近实际非线性的函数更好并且在噪音赔偿和扬声器改编,MLPR的词错误率是比MLLR的那些显著地低的。建议MLPR算法克服好的线性假设和罐头减少的限制噪音,说话者和另外的因素的影响同时。它对说话者和噪音的联合改编特别合适。