关于数值计算方法收敛性的一点认识

收敛性是数值计算方法中一个非常重要的概念.采用各种数值计算方法求解了常微分方程初值问题,试图通过哲学公式相对真理/绝对真理=0.9来解释数值计算结果和理论结果的关系.通过此哲学公式来刻画数值解收敛到真解的过程,简单易懂.随着小数点后面9的个数的增加,数值结果和理论结果的误差在不断减小.哲学公式有助于学生进一步认识数值计算方法的收敛性.     

计算半线性椭圆问题多解的一类谱Galerkin型搜索延拓法的收敛性分析

本文提出计算半线性椭圆边值问题多解的一类高效的谱Galerkin型搜索延拓法(SGSEM).该方法基于模型方程相应线性特征值问题的若干特征函数的线性组合构造多解初值,充分利用了传统搜索延拓法构造多解初值方面的优势.同时,采用插值系数Legendre-Galerkin谱方法离散模型问题,具有计算成本低、计算精度高的优点.运用Schauder不动点定理和其他技巧,本文严格证明了对应于每个特定真解的数值解的存在性以及限制在该真解一个充分小的邻域内的数值解的唯一性,并证明了其谱收敛性.数值结果验证了算法的可行性与高效性,并展示了不同类型的多解.     

外推瀑布多网格法(EXCMG)——-大规模求解椭圆问题的新算法

基于有限元的渐近展开式,导出了新的外推公式,它们更精确地逼近密网上的有限元解(而不是微分方程的解).提出了新的外推瀑布型多网格法(EXCMG),采用新外推公式及其二次插值提供密网上的好初值.数值实验表明,新方法有很高的精度和效率.最后在PC机上求解了大规模二维椭圆问题.     

Plate Contact问题的混合有限元逼近

论文考虑了Plate Contact问题的混合有限元逼近,其变分问题为第二类四阶椭圆变分不等问题.首先,根据正则化方法,得到原问题的正则化问题.再根据网格依赖范数技巧,考虑了正则化问题的Ciarlet-Raviart混合有限元逼近,并证明了真解与逼近解之间的误差估计.最后通过数值算例验证了理论分析的结果.     

解Stokes特征值问题的一种两水平稳定化有限元方法

基于局部Gauss积分,研究了解Stokes特征值问题的一种两水平稳定化有限元方法．该方法涉及在网格步长为H的粗网格上解一个Stokes特征值问题,在网格步长为h=O(H2)的细网格上解一个Stokes问题．这样使其能够仍旧保持最优的逼近精度,求得的解和一般的稳定化有限元解具有相同的收敛阶,即直接在网格步长为h的细网格上解一个Stokes特征值问题．因此,该方法能够节省大量的计算时间．数值试验验证了理论结果．     

对流扩散方程特征线三角元法的一致估计

利用三角形线性元的积分恒等式,给出了二维非定常对流占优扩散方程的特征线有限元解和真解的一致最优估计,并利用插值后处理算子,得到了有限元解梯度的一致超收敛估计,即只与初值和右端项有关,而与ε无关.     

对流扩散方程特征线双线性元的一致估计

利用双线性元的积分恒等式,给出了二维非定常对流占优扩散方程的特征线有限元解和真解的一致误差估计,并利用插值后处理算子给出了有限元解梯度的一致超收敛估计,即上述误差与ε无关,而仅与右端f和初值u_0有关.     

对非线性方程近似解的认识和思考

二分法和牛顿法求非线性方程根的近似值已列入中学课程.但它背后的哲学原理(相对真理)/(绝对真理)=0.9,只在林群的新书中说到2^(1/2)时提出来.根据教学需要,通过(不足近似值)/(过剩近似值)=0.9等数值化的公式,来刻画根的近似过程.可以清楚地看到,随着小数点后9的个数的增加,近似解和真实解的误差在不断减小.因此0.9数值化系列公式也可以看做是误差估计的另一种表型形式.     

有限元方法(续完)

§3. 变分问题的离散化 本节以变分问题(1.25)为例说明在三角剖分线性插值基础上的离散化过程和一些情况的处理,见§§3.1—5,但不涉及离散化后代数方程的数值解法,对此可参考后引的著作.§3.6讨论离散解对于真解的收敛性.§3.7对有限元方法作一评价。 为了说明一般的原则,不妨把问题(1.25)稍为推广如下:     

对流扩散方程三角形有限元解的一致估计

利用三角形线性元的积分恒等式,给出了二维非定常对流扩散方程的半离散有限元解和真解的一致最优误差估计,即误差与ε无关,而仅与右端f和初值u_0有关.     

关于一些概率统计理论的一点认识

近来,林群从哲学的角度用公式相对真理/绝对真理=0.9阐述了微积分的内涵,并探究了计算数学、概率统计等学科领域乃至日常生活现象也同样蕴藏着该哲学公式.基于林群等在《概率论初步设想》中对大数定律和中心极限定理等概率统计学理论的哲学内涵分析,用一些数值算例展示概率统计学中这些常用理论所蕴含的这一哲学公式.     

Numerical Solutions for second-kind Volterra integral equations by Galerkin methods

In this paper, we study the global convergence for the numerical solutions of nonlinear Volterra integral equations of the second kind by means of Galerkin finite element methods. Global superconvergence properties are discussed by iterated finite element methods and interpolated finite element methods. Local superconvergence and iterative correction schemes are also considered by iterated finite element methods. We improve the corresponding results obtained by collocation methods in the recent papers [6] and [9] by H. Brunner, Q. Lin and N. Yan. Moreover, using an interpolation post-processing technique, we obtain a global superconvergence of the O(h ^2r )-convergence rate in the piecewise-polynomial space of degree not exceeding (r–1). As a by-product of our results, all these higher order numerical methods can also provide an a posteriori error estimator, which gives critical and useful information in the code development.     

Superconvergence of splitting positive definite mixed finite element for parabolic optimal control problems

In this paper, we investigate the superconvergence of fully discrete splitting positive definite mixed finite element (MFE) methods for parabolic optimal control problems. For the space discretization, the state and co-state are approximated by the lowest order Raviart–Thomas MFE spaces and the control variable is approximated by piecewise constant functions. The time discretization of the state and co-state are based on finite difference methods. We derive the superconvergence between the projections of exact solutions and numerical solutions or the exact solutions and postprocessing numerical solutions for the control, state and co-state. A numerical example is provided to validate the theoretical results.     

Analysis of some mixed elements for the Stokes problem

In this paper we discuss some mixed finite element methods related to the reduced integration penalty method for solving the Stokes problem. We prove optimal order error estimates for bilinear-constant and biquadratic-bilinear velocity-pressure finite element solutions. The result for the biquadratic-bilinear element is new, while that for the bilinear-constant element improves the convergence analysis of Johnson and Pitkäranta (1982). In the degenerate case when the penalty parameter is set to be zero, our results reduce to some related known results proved in by Brezzi and Fortin (1991) for the bilinear-constant element, and Bercovier and Pironneau (1979) for the biquadratic-bilinear element. Our theoretical results are consistent with the numerical results reported by Carey and Krishnan (1982) and Oden et al. (1982).     

无穷凹角区域各向异性问题的自然边界元法

本文研究无穷凹角区域上一类各向异性问题的自然边界元法.利用自然边界归化原理,获得该问题的Poisson积分公式和自然积分方程,给出了自然积分方程的数值方法,以及逼近解的收敛性和误差估计,最后给出了数值例子,以示方法的可行性和有效性.     

High-order dual-parametric finite element methods for cavitation computation in nonlinear elasticity

In this paper, we present the numerical analysis on high order dual parametric finite element methods for the cavitation computation problems in nonlinear elasticity, which leads to a meshing strategy assuring high efficiency on numerical approximations to cavity deformations. Furthermore, to cope with the high order approximation of the finite element methods, properly chosen weighted Gaussian type numerical quadrature is applied to the singular part of the elastic energy. Our numerical experiments show that the high order dual parametric finite element methods work well when coupled with properly designed weighted Gaussian type numerical quadratures for the singular part of the elastic energy, and the convergence rates of the numerical cavity solutions are shown to be significantly improved as expected.     

A Posteriori Error Estimates of Mixed Methods for Parabolic Optimal Control Problems

In this article, we shall give a brief review on the fully discrete mixed finite element method for general optimal control problems governed by parabolic equations. The state and the co-state are approximated by the lowest order Raviart–Thomas mixed finite element spaces and the control is approximated by piecewise constant elements. Furthermore, we derive a posteriori error estimates for the finite element approximation solutions of optimal control problems. Some numerical examples are given to demonstrate our theoretical results.     

Galerkin finite element methods for two-dimensional RLW and SRLW equations

In this paper, Galerkin finite methods for two-dimensional regularized long wave and symmetric regularized long wave equation are studied. The discretization in space is by Galerkin finite element method and in time is based on linearized backward Euler formula and extrapolated Crank–Nicolson scheme. Existence and uniqueness of the numerical solutions have been shown by Brouwer fixed point theorem. The error estimates of linearlized Crank–Nicolson method for RLW and SRLW equations are also presented. Numerical experiments, including the error norms and conservation variables, verify the efficiency and accuracy of the proposed numerical schemes.     

Improved convergence theorems of Newton's method designed for the numerical verification for solutions of differential equations

Though the convergence theorem of simplified Newton's method is an excellent general principle for the numerical verification of isolated solutions of differential equations, it is not always good from the viewpoint of computational efficiency, in particular when we use finite element solutions as approximate solutions. We improve the theorem to overcome this point. Some numerical examples on the nonlinear elliptic equations show that the remarkable increase of computational efficiency is achieved by our improvement.     

Convergence of multi-level iterative aggregation-disaggregation methods

This paper introduces an error propagation formula of a certain class of multi-level iterative aggregation-disaggregation (IAD) methods for numerical solutions of stationary probability vectors of discrete finite Markov chains. The formula can be used to investigate convergence by computing the spectral radius of the error propagation matrix for specific Markov chains. Numerical experiments indicate that the same type of the formula could be used for a wider class of the multi-level IAD methods. Using the formula we show that for given data there is no relation between convergence of two-level and of multi-level IAD methods.