Relaxed Steepest Descent and Cauchy-Barzilai-Borwein Method

The negative gradient direction to find local minimizers has been associated with the classical steepest descent method which behaves poorly except for very well conditioned problems. We stress out that the poor behavior of the steepest descent methods is due to the optimal Cauchy choice of steplength and not to the choice of the search direction. We discuss over and under relaxation of the optimal steplength. In fact, we study and extend recent nonmonotone choices of steplength that significantly enhance the behavior of the method. For a new particular case (Cauchy-Barzilai-Borwein method), we present a convergence analysis and encouraging numerical results to illustrate the advantages of using nonmonotone overrelaxations of the gradient method.     

Rayleigh quotient algorithms for nonsymmetric matrix pencils

A classical Rayleigh-quotient iterative algorithm (known as “broken iteration”) for finding eigenvalues and eigenvectors is applied to semisimple regular matrix pencils A − λB. It is proved that cubic convergence is attained for eigenvalues and superlinear convergence of order three for eigenvectors. Also, each eigenvalue has a local basin of attraction. A closely related Newton algorithm is examined. Numerical examples are included. Dedicated to the memory of Gene H. Golub.     

Asymptotic Convergence of Conjugate Gradient Methods for the Partial Symmetric Eigenproblem

Recently an efficient method (DACG) for the partial solution of the symmetric generalized eigenproblem A x = δB x has been developed, based on the conjugate gradient (CG) minimization of the Rayleigh quotient over successive deflated subspaces of decreasing size. The present paper provides a numerical analysis of the asymptotic convergence rate ρ_j of DACG in the calculation of the eigenpair λ_j, u _j, when the scheme is preconditioned with A⁻¹. It is shown that, when the search direction are A-conjugate, ρ_j is well approximated by 4/ξ_j, where ξ_j is the Hessian condition number of a Rayleigh quotient defined in appropriate oblique complements of the space spanned by the leftmost eigenvectors u ₁, u ₂,…, u _j−1 already calculated. It is also shown that 1/ξ_j is equal to the relative separation between the eigenvalue λ_j currently sought and the next higher one λ_j+1 and the next higher one λ_{j + 1}. A modification of DACG (MDACG) is studied, which involves a new set of CG search directions which are made M-conjugate, with M-conjugate, with M-conjugate, with M a matrix approximating the Hessian. By distinction, MDACG has an asymptotic rate of convergence which appears to be inversely proportional to the square root of ξ_j, in complete agreement with the theoretical results known for the CG solution to linear systems. © 1997 by John Wiley & Sons, Ltd.     

空间型中紧致无边子流形上的第一特征值

得到了两个关于空间形式中紧致无边子流形的广义位置向量场和其上Laplace算子第一特征值λ_1的积分不等式。并由此首先给出了λ_1与其上界间的间隔估计,其次得到了此紧致无边子流形等距浸入在空间形式的测地超球面或等距于测地超球面的充分条件,推广了Deshmukh[6]在欧氏空间中的相应结论。     

On the best constant for Hardy's inequality in

Let be a domain in and . We consider the (generalized) Hardy inequality , where . The inequality is valid for a large family of domains, including all bounded domains with Lipschitz boundary. We here explore the connection between the value of the Hardy constant and the existence of a minimizer for this Rayleigh quotient. It is shown that for all smooth -dimensional domains, , where is the one-dimensional Hardy constant. Moreover it is shown that for all those domains not possessing a minimizer for the above Rayleigh quotient. Finally, for , it is proved that if and only if the Rayleigh quotient possesses a minimizer. Examples show that strict inequality may occur even for bounded smooth domains, but for convex domains.

     

Accuracy of Computed Eigenvectors Via Optimizing a Rayleigh Quotient

This paper establishes converses to the well-known result: for any vector such that the sine of the angle sin(u, )=O(), we have

ON THE RAYLEIGH QUOTIENT FOR SINGULAR VALUES

In this paper, the theoretical analysis for the Rayleigh quotient matrix is studied, some results of the Rayleigh quotient （matrix） of Hermitian matrices are extended to those for arbitrary matrix on one hand. On the other hand, some unitarily invariant norm bounds for singular values are presented for Rayleigh quotient matrices. Our results improve the existing bounds.     

基于辛本征空间的线性阻尼振动系统动力学分析

对于考虑阻尼项和陀螺项的一般线性动力学振动系统,建立基于辛本征空间展开求解的一般方法.基于Rayleigh商本征值的模态展开方法被广泛应用于复杂结构动力系统振动分析,但对于很多机械系统,由于其不能有效考虑陀螺效应的影响,其适用性却受到很大限制.该文首先讨论了无阻尼系统Rayleigh商本征值问题与辛本征值问题的对应关系,表明前者实际可由后者的一种退化形式给出（也即忽略陀螺效应）,而后者更具有一般性.在此基础上,进一步基于辛本征空间本征向量展开,推导了同时考虑阻尼和陀螺系统的一般线性动力学系统的有效求解方法.数值算例选取不考虑陀螺效应及考虑陀螺效应的两种线性阻尼振动系统对所提出的方法进行了验证,分析结果表明了该文所建立方法的正确性和有效性.     

膜与不可压缩有界流体接触时的自由振动

圆形膜与流体接触时的振动被广泛应用于工业中．推导了圆形膜在与不可压缩有界流体接触时,非对称自由振动的自振频率．鉴于膜在不可压缩、非粘性流体中振动引起的小振幅,采用速度势函数来描述流体场．使用两种方法来推导系统的自由振动频率．它们包括变分法及一种近似解法——Rayleigh商法．在用两种方法求得的自由振动频率值之间具有良好的相关性．最后,研究了流体的深度、质量密度以及径向张力对耦合系统自由振动频率的影响．     

New schemes with fractal error compensation for PDE eigenvalue computations

With an error compensation term in the fractal Rayleigh quotient of PDE eigen-problems,we propose a new scheme by perturbing the mass matrix Mhto Mh=Mh+Ch2mKh,where Khis the corresponding stif matrix of a 2m 1 degree conforming finite element with mesh size h for a 2m-order self-adjoint PDE,and the constant C exists in the priority error estimationλh jλj~Ch2mλ2j.In particular,for Laplace eigenproblems over regular domains in uniform mesh,e.g.,cube,equilateral triangle and regular hexagon,etc.,we find the constant C=I h 1Mh2 hKh and show that in this case the computation accuracy can raise two orders,i.e.,fromλh jλj=O(h2)to O(h4).Some numerical tests in 2-D and 3-D are given to verify the above arguments.     

一般边界条件下Sturm-Liouville算子的Ambarzumyan型定理

该文研究有限区间上一般自伴边界条件下的Sturm-Liouville方程的逆特征值问题.将Neumann边界条件下Sturm-Liouville方程的Ambarzumyan型定理推广到一般自伴边界条件下情形,证明了如果它的特征值与零势的特征值一样,则Sturm-Liouville方程的势为零.     

复合材料特征值的高阶多尺度Rayleigh商校正

本文讨论了周期结构复合材料特征值的多尺度计算,提出了高阶多尺度Rayleigh商校正算法,并给出了收敛性分析. 最后,通过大量数值实验结果表明,新算法是有效且必要的.     

Inexact Rayleigh Quotient-Type Methods for Eigenvalue Computations

We consider the computation of an eigenvalue and corresponding eigenvector of a Hermitian positive definite matrix A , assuming that good approximations of the wanted eigenpair are already available, as may be the case in applications such as structural mechanics. We analyze efficient implementations of inexact Rayleigh quotient-type methods, which involve the approximate solution of a linear system at each iteration by means of the Conjugate Residuals method. We show that the inexact version of the classical Rayleigh quotient iteration is mathematically equivalent to a Newton approach. New insightful bounds relating the inner and outer recurrences are derived. In particular, we show that even if in the inner iterations the norm of the residual for the linear system decreases very slowly, the eigenvalue residual is reduced substantially. Based on the theoretical results, we examine stopping criteria for the inner iteration. We also discuss and motivate a preconditioning strategy for the inner iteration in order to further accelerate the convergence. Numerical experiments illustrate the analysis.     

Homotopy method for the numerical solution of the eigenvalue problem of self-adjoint partial differential operators

GivenA ₁, the discrete approximation of a linear self-adjoint partial differential operator, the smallest few eigenvalues and eigenvectors ofA ₁ are computed by the homotopy (continuation) method. The idea of the method is very simple. From some initial operatorA ₀ with known eigenvalues and eigenvectors, define the homotopyH(t)=(1–t)A ₀+tA₁, 0t1. If the eigenvectors ofH(t ₀) are known, then they are used to determine the eigenpairs ofH(t ₀+dt) via the Rayleigh quotient iteration, for some value ofdt. This is repeated untilt becomes 1, when the solution to the original problem is found. A fundamental problem is the selection of the step sizedt. A simple criterion to selectdt is given. It is shown that the iterative solver used to find the eigenvector at each step can be stabilized by applying a low-rank perturbation to the relevant matrix. By carrying out a small part of the calculation in higher precision, it is demonstrated that eigenvectors corresponding to clustered eigenvalues can be computed to high accuracy. Some numerical results for the Schrödinger eigenvalue problem are given. This algorithm will also be used to compute the bifurcation point of a parametrized partial differential equation.Dedicated to Herbert Bishop Keller on the occasion of his 70th birthdayThe work of this author was in part supported by RGC Grant DAG93/94.SC30.The work of this author was in part supported by NSF Grant DMS-9403899.