简介:无单元伽辽金(EFG)法采用移动最小二乘近似构造形函数,从能量泛函的变分形式出发得到控制方程,并用罚函数法施加本质边界条件,从而得到偏微分边值问题的数值解.改进的广义移动最小二乘近似(IGMLS)在构造函数时要求近似函数在所有节点处误差的平方和与近似函数导数仅在导数边界附近各节点处误差的平方和之和最小.同时,为了节省计算时间,基函数采用加权正交多项式.将IGMLS与EFG相结合,对板弯曲离散建立了相应的代数方程.通过数值算例证实了IGMKS比改进的移动最小二乘近似(IMLS)具有更高的精度,所需的运算时间要小于广义移动最小二乘近似(GMLS).
简介:摘要:采用无单元 Galerkin 方法数值求解具有狄利克雷边界条件的二维瞬态热传导问题。首先离散该问题的时间变量,将该问题 转化为与时间无关的边值问题;然后采用罚函数法处理 Dirichlet 边界条件,得到数值离散方程组,再利用 Matlab 软件求解给出的算例,结果表明该方法得到的数值结果与解析解吻合较好,该方法具有较高的计算精度和较好的收敛性。
简介:Abstract.OgrobjectinthisartlcleistodescribetbeGalerklnschemeandnonlin-eaxGalerkinschemefortheapproximationofnonlinearevolutionequations,andtostudythestabilityoftheseschemes.SpatialdiscretizatloncanbepedormedbyeitherGalerklnspectralmethodornonlinearGalerldnspectralmethod;timediscretizatlortisdonehyEulersin.heinewklchisexplicitorimplicitinthenonlinearterms.Accordingtothestabilityanalysisoftheaboveschemes,thestabilityofnonllneexGalerklnmethodisbetterthanthatofGalexklnmethod.
简介:IntroductionTheconceptsofInertialManifold(IM)[1]andApproximateInertialManifold(AIM)[2]fordissipativepartialdifferentialequati...
简介:TheGalerkinandleast-squaresmethodsaretwoclassesofthemostpopularKrylovsubspacemethOdsforsolvinglargelinearsystemsofequations.Unfortunately,boththemethodsmaysufferfromseriousbreakdownsofthesametype:InabreakdownsituationtheGalerkinmethodisunabletocalculateanapproximatesolution,whiletheleast-squaresmethod,althoughdoesnotreallybreakdown,isunsucessfulinreducingthenormofitsresidual.Inthispaperwefrstestablishaunifiedtheoremwhichgivesarelationshipbetweenbreakdownsinthetwometh-ods.Wefurtherillustratetheoreticallyandexperimentallythatifthecoefficientmatrixofalienarsystemisofhighdefectivenesswiththeassociatedeigenvalueslessthan1,thentherestart-edGalerkinandleast-squaresmethodswillbeingreatrisksofcompletebreakdowns.Itappearsthatourfindingsmayhelptounderstandphenomenaobservedpracticallyandtoderivetreat-mentsforbreakdownsofthistype.
简介:Inthispaper,weareconcernedwithuniformsuperconvergenceofGalerkinmethodsforsingularlyperturbedreaction-diffusionproblemsbyusingtwoShishkin-typemeshes.Basedonanestimateoftheerrorbetweensplineinterpolationoftheexactsolutionanditsnumericalapproximation,aninterpolationpost-processingtechniqueisappliedtotheoriginalnumericalsolution.Thisresultsinapproximationexhibitsuperconvergencewhichisuniformintheweightedenergynorm.Numericalexamplesarepresentedtodemonstratetheeffectivenessoftheinterpolationpost-processingtechniqueandtoverifythetheoreticalresultsobtainedinthispaper.
简介:Inthispaper,theminimaldissipationlocaldiscontinuousGalerkinmethodisstudiedtosolvetheellipticinterfaceproblemsintwo-dimensionaldomains.Theinterfacemaybearbitrarysmoothcurves.ItisshownthattheerrorestimatesinL2-normforthesolutionandthefluxareO(h2|logh|)andO(h|logh|^l/2),respectively.Innumericalexperiments,thesuccessivesubstitutioniterativemethodsareusedtosolvetheLDGschemes.Numericalresultsverifytheefficiencvandaccuracvofthemethod.
简介:ItisprovedinthispaperthattheapproximatesolutionofthediscontinuousGalerkinmethoddoesconvergeeventheexactsolutionofthefirstorderhyperbolicequationisdiscontinuous.
简介:AD(Alternatingdirection)Galerkinschemesford-dimensionalnonlinearpseudo-hyperbolicequationsarestudied.Byusingpatchapproximationtechnique,ADprocedureisrealized,andcalculation,workissimplified.ByusingGalerkinapproach,highlycomputationalaccuracyiskept.Byusingvariousprioriestimatetechniquesfordifferentialequations,difficultycomingformnon-linearityistreated,andoptimalH^1andL^2convergenceprop-ertiesaredemonstrated.Moreover,althoughalltheexistedADGalerkinschemesusingpatchapproximationarelimitedtohaveonlyoneorderaccuracyintimeincrement,yettheschemesformulatedinthispaperhavesecondorderaccuracyinit.ThisimpliesanessentialadvancementinADGalerkinaualysis.
简介:Akindofcalculatingmethodforhighorderdifferentialsexpandedbythewaveletscal-ingfunctionsandtheintegraloftheirproductusedinGalerkinFEMisproposed,sothatwecanusethewaveletGalerkinFEMtosolveboundary-valuedifferentialequationswithordershigherthantwo.TocombinethismethodwiththeGeneralizedGaussianintegralmethodinwavelettheory,wecanfindthatthismethodhasmanymeritsinsolvingmechanicalproblems,suchasthebendingofplatesandthosewithvariablethickness.Thenumericalresultsshowthatthismethodisaccurate.