大型线性方程组的变分迭代解法

本文提出了一种求解大型线性方程组的一种新方法——变分迭代解法．这种方法的基本思想是：先给方程一个近似的初值,然后引进若干个拉氏乘子控正其近似值．而拉氏乘于可用垭值的概念最佳确定．这种方法收敛速度较快,如果只取,n个拉氏乘子(n为方程个数),则该方法即为Newton迭代法．     

Schrodinger方程的变分迭代解法

该文以Ｓｃｈｒｏｄｉｎｇｅｒ方程为例,分析变分迭代法的一些基本特点．在该方法中引进了一广义拉氏乘子构造了一迭代格式,拉氏乘子可由变分理论最佳识别．由于在识别拉氏乘子是应用了限制变分的概念,所以只能通过迭代才能得到收敛解．为了加快收敛速度,可以在初始近似引入待定常数,而待定常数又可用各种方式最佳识别．文中初步分析了该方法的收敛性,对于Ｓｃｈｒｏｄｉｎｇｅｒ方程,其一阶近似即可得到Ｊｏｓｔ解．     

改进的拉氏乘子法和多变量广义变分原理

拉氏乘子法是构造广义变分原理的重要途径 ,在识别拉氏乘子时 ,拉氏乘子是独立变分的 ,而识别后 ,它却是其他变量的函数 ,这是产生临界变分的原因 .本文对拉氏乘子法作了改进 ,提出了一种新的理论——凑合反推法 ,应用该方法可以方便地构造多变量的广义变分原理 ,并且不会出现临界变分现象     

高阶拉氏乘子法和弹性理论中更一般的广义变分原理

作者曾指出[1],弹性理论的最小位能原理和最小余能原理都是有约束条件限制下的变分原理采用拉格朗日乘子法,我们可以把这些约束条件乘上待定的拉氏乘子,计入有关变分原理的泛函内,从而将这些有约束条件的极值变分原理,化为无条件的驻值变分原理.如果把这些待定拉氏乘子和原来的变量都看作是独立变量而进行变分,则从有关泛函的驻值条件就可以求得这些拉氏乘子用原有物理变量表示的表达式.把这些表达式代入待定的拉氏乘子中,即可求所谓广义变分原理的驻值变分泛函.但是某些情况下,待定的拉氏乘子在变分中证明恒等于零.这是一种临界的变分状态.在这种临界状态中,我们无法用待定拉氏乘子法把变分约束条件吸收入泛函,从而解除这个约束条件.从最小余能原理出发,利用待定拉氏乘子法,企图把应力应变关系这个约束条件吸收入有关泛函时,就发生这种临界状态,用拉氏乘子法,从余能原理只能导出Hellinger-Reissner变分原理[2],[3],这个原理中只有应力和位移两类独立变量,而应力应变关系则仍是变分约束条件,人们利用这个条件,从变分求得的应力中求应变.所以Hellinger-Reissner变分原理仍是一种有条件的变分原理.     

对弹性理论中临界变分状态的一个注记

(t) 最近钱伟长教授指出[1],在某些情况下,用普通的拉氏乘子法,其待定的拉氏乘子在变分中恒等于零,这称为临界变分状态,在这种临界状态中,我们无法用待定拉氏乘子法把变分的约束条件吸收入泛函,从而解除这个约束条件.例如用拉氏乘子法,从最小余能原理只能导出Hellinger-Reissner变分原理,这个原理中只有应力和位移两类独立变量,而应力应变关系仍然是变分的约束条件.为了消除这个约束条件,钱伟长教授提出了高次拉氏乘子法,即在泛函中引入二次项
A_ifk1(e_ij-b_iimnσmn)(e_ki-b_k1pqσ_pq)
来消除应力应变这个约束条件. 本文目的是要证明,如果在泛函中引入如下二次项
A_ifk1(e_ij-b_iimnσ_mn)(e_ki-1/2u_k2-1/2u_1:k)
我们也可以用高次拉氏乘子法解除应力应变这个变分约束条件.用这种方法,我们不仅可以从Hel-linger-Reissner原理的基础上,找到更一般的广义变分原理.在特殊情况下,这个更一般的广义变分原理,可以还原为各种已知的弹性理论变分原理.同样,我们也可以从Hu-Washizu(胡海昌-鹫津久一郎)[4,5]变分原理,用高次拉氏乘子法,求得比该原理更一般的广义变分原理.     

分数阶Fornberg-Whitham方程及其改进方程的变分迭代解

给出分数阶Fornberg Whitham方程(FFW)并把其中非线性项uu_x换为u2u_x后所得的改进Fornberg-Whitham方程的解．使用了分数阶变分迭代法(fractional variational iteration method,FVIM),其中Lagrange乘子由泛函和Laplace变换确定．讨论了分数阶次的数值在两种情况下FFW方程的解,因为确定FFW方程中时间微分的阶次需要比较原方程中含时间的两个微分的阶次．最后,给出两个使用分数阶变分迭代法的算例．算例结果证明了所提方法的有效性     

对合变换和薄板弯曲问题的多变量变分原理

本文利用拉氏乘子法把薄板弯曲问题的最小位能原理和最小余能原理的变分约束条件解除.从而导出了常见的广义变分原理.为了降低泛函中变量导数的阶次.我们用对合变换引进新的正则变量.于是,我们可以进一步利用拉氏乘子法,把这些对合变换当作变分约束而予以消除,从而导出了各种多变量的薄板弯曲广义变分原理.事实证明,使用上述拉氏乘子法,并不能消除一切变分约束;为此,我们进一步引用高阶拉氏乘子法消除这些剩下来的约束条件,从而导得了薄板弯曲问题的更一般的广义变分原理.     

基于几何非协调分解的Lagrange乘子区域分解方法

本文考虑将Lagrange乘子区域分解方法应用于几何非协调分解的情况来求解二阶椭圆问题.由于采用几何非协调区域分解,每个局部乘子空间关联到多个界面,我们按照一定的规则选取合适的乘子面来定义乘子空间.利用局部正则化技巧,可以消去内部变量,得到关于Lagrange乘子的界面方程.采用一种经济的预条件迭代方法求解界面方程,且相关的预条件子是可扩展的.     

极值曲面和非线性偏微分方程的一般解

本文得到一类二阶非线性偏微分方程的通解公式，这类方程与ｎ维欧氏空间Ｒｎ或ｎ维Ｍｉｎｋｏｗｓｋｉ空间Ｒ中的极值曲面密切相关．     

另一类非完整力学系统的Lagrange方程

用文[1]的方法,导出另一类一阶非完整力学系统不带乘子的Lagrange方程.这种形式的方程也是新的.     

A multilevel correction method for Stokes eigenvalue problems and its applications

In this paper, a new multilevel correction scheme is proposed to solve Stokes eigenvalue problems by the finite element method. This new scheme contains a series of correction steps, and the accuracy of eigenpair approximation can be improved after each step. In each correction step, we only need to solve a Stokes problem on the corresponding fine finite element space and a Stokes eigenvalue problem on the coarsest finite element space. This correction scheme can improve the efficiency of solving Stokes eigenvalue problems by the finite element method. As applications of this multilevel correction method, a multigrid method and an adaptive finite element technique are introduced for Stokes eigenvalue problems. Some numerical results are given to validate our schemes. Copyright © 2013 John Wiley & Sons, Ltd.     

A Multilevel Correction Method for Steklov Eigenvalue Problem by Nonconforming Finite Element Methods

In this paper, a multilevel correction scheme is proposed to solve the Steklov eigenvalue problem by nonconforming finite element methods. With this new scheme, the accuracy of eigenpair approximations can be improved after each correction step which only needs to solve a source problem on finer finite element space and an Steklov eigenvalue problem on the coarsest finite element space. This correction scheme can increase the overall efficiency of solving eigenvalue problems by the nonconforming finite element method. Furthermore, as same as the direct eigenvalue solving by nonconforming finite element methods, this multilevel correction method can also produce the lower-bound approximations of the eigenvalues.     

Alternative correction equations in the Jacobi‐Davidson method

The correction equation in the Jacobi‐Davidson method is effective in a subspace orthogonal to the current eigenvector approximation, whereas for the continuation of the process only vectors orthogonal to the search subspace are of importance. Such a vector is obtained by orthogonalizing the (approximate) solution of the correction equation against the search subspace. As an alternative, a variant of the correction equation can be formulated that is restricted to the subspace orthogonal to the current search subspace. In this paper, we discuss the effectiveness of this variant. Our investigation is also motivated by the fact that the restricted correction equation can be used for avoiding stagnation in the case of defective eigenvalues. Moreover, this equation plays a key role in the inexact TRQ method [18]. Copyright © 1999 John Wiley & Sons, Ltd.     

Optimal Error Correction and Methods of Feasible Directions

The main objective of this study is to discuss the optimum correction of linear inequality systems and absolute value equations (AVE). In this work, a simple and efficient feasible direction method will be provided for solving two fractional nonconvex minimization problems that result from the optimal correction of a linear system. We will show that, in some special-but frequently encountered-cases, we can solve convex optimization problems instead of not-necessarily-convex fractional problems. And, by using the method of feasible directions, we solve the optimal correction problem. Some examples are provided to illustrate the efficiency and validity of the proposed method.     

A new approach to nonlinear partial differential equations

In this paper, a novel method called variational iteration method is proposed to solve nonlinear partial differential equations without linearization or small perturbations. In this method, a correction functional is constructed by a general Lagrange multiplier, which can be identified via variational theory. An analytical solution can be obtained from its trial-function with possible unknown constants, which can be identified by imposing the boundary conditions, by successively iteration.     

Improvements of Some Projection Methods for Monotone Nonlinear Variational Inequalities

In this paper, we study the relationship of some projection-type methods for monotone nonlinear variational inequalities and investigate some improvements. If we refer to the Goldstein–Levitin–Polyak projection method as the explicit method, then the proximal point method is the corresponding implicit method. Consequently, the Korpelevich extragradient method can be viewed as a prediction-correction method, which uses the explicit method in the prediction step and the implicit method in the correction step. Based on the analysis in this paper, we propose a modified prediction-correction method by using better prediction and correction stepsizes. Preliminary numerical experiments indicate that the improvements are significant.     

Application of He’s variational iteration method to Helmholtz equation

In this article, we implement a new analytical technique, He’s variational iteration method for solving the linear Helmholtz partial differential equation. In this method, general Lagrange multipliers are introduced to construct correction functionals for the problems. The multipliers in the functionals can be identified optimally via the variational theory. The initial approximations can be freely chosen with possible unknown constants, which can be determined by imposing the boundary/initial conditions. The results compare well with those obtained by the Adomian’s decomposition method.     

LOCAL PRESSURE CORRECTION FOR THE STOKES SYSTEM

Pressure correction methods constitute the most widely used solvers for the timedependent Navier-Stokes equations.There are several different pressure correction methods,where each time step usually consists in a predictor step for a non-divergence-free velocity,followed by a Poisson problem for the pressure(or pressure update),and a final velocity correction to obtain a divergence-free vector field.In some situations,the equations for the velocities are solved explicitly,so that the numerical most expensive step is the elliptic pressure problem.We here propose to solve this Poisson problem by a domain decomposition method which does not need any communication between the sub-regions.Hence,this system is perfectly adapted for parallel computation.We show under certain assumptions that this new scheme has the same order of convergence as the original pressure correction scheme(with global projection).Numerical examples for the Stokes system show the effectivity of this new pressure correction method.The convergence order O(k^2)for resulting velocity fields can be observed in the norm l^2(0,T;L^2(Ω)).     

Bayes factors in the presence of population stratification

Hidden population stratification (PS) is a main concern in the analysis of case-control genetic association studies. All methods to correct for hidden PS have been focused on classical hypothesis testing, and cannot be directly applied to Bayesian analysis. In this paper, to study the impact and the correction of hidden PS on Bayes factor (BF), we use a simple approximation of BF in terms of the maximum likelihood estimates of the odds ratio (OR) and its asymptotic variance. One advantage is that the commonly used principal components analysis method with a large panel of null markers scanned from existing genome-wide association studies can be directly employed to correct for hidden PS in estimating the OR and its asymptotic variance, through which a correction to BF for hidden PS can be achieved. Using simulations, we examine the impact of ignoring hidden PS on BF and show that the proposed method yields an appropriate correction in Bayesian analysis.     

A Jacobi-Davidson method for two real parameter nonlinear eigenvalue problems arising from delay differential equations

The critical delays of a delay-differential equation can be computed by solving a nonlinear two-parameter eigenvalue problem. For large scale problems, we propose new correction equations for a Jacobi-Davidson type method, that also forces real valued critical delays. We present two different equations: one complex valued equation using a direct linear system solver, and one Jacobi-Davidson style correction equation which is suitable for an iterative linear system solver. A numerical example of a large scale problem arising from PDEs shows the effectiveness of the method. (© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim)