A MODIFIED INVERSE ITERATION FOR A LARGE SPARSE SPD GENERALIZED EIGENPROBLEM

In this paper, an algorithm based on a shifted inverse power iteration for computing generalized eigenvalues with corresponding eigenvectors of a large scale sparse symmetric positive definite matrix pencil is presented. It converges globally with a cubic asymptotic convergence rate, preserves sparsity of the original matrices and is fully parallelizable. The algebraic multilevel itera-tion method (AMLI) is used to improve the efficiency when symmetric positive definite linear equa-tions need to be solved.     

Increasing efficiency of inverse iteration

Inverse iteration is simple but not very efficient method for computing few eigenvalues with minimal absolute values and corresponding eigenvectors of a symmetric matrix. The idea is to increase its efficiency by technique similar to multigrid methods used for solving linear systems. This approach is not new, but until now multigrid was mostly used for solving linear system which appear in Rayleigh quotient iteration, inverse iteration and related iterative methods. Instead of choosing appropriate coordinates (grids), our algorithm performs inverse iteration on a sequence of subspaces with decreasing dimensions (multispace). Block Lanczos method is used for the selection of a smaller subspace. This will produce a banded matrix, which makes inverse iteration even faster in the smaller dimensions.      

A geometric theory for preconditioned inverse iteration applied to a subspace

The aim of this paper is to provide a convergence analysis for a preconditioned subspace iteration, which is designated to determine a modest number of the smallest eigenvalues and its corresponding invariant subspace of eigenvectors of a large, symmetric positive definite matrix. The algorithm is built upon a subspace implementation of preconditioned inverse iteration, i.e., the well-known inverse iteration procedure, where the associated system of linear equations is solved approximately by using a preconditioner. This step is followed by a Rayleigh-Ritz projection so that preconditioned inverse iteration is always applied to the Ritz vectors of the actual subspace of approximate eigenvectors. The given theory provides sharp convergence estimates for the Ritz values and is mainly built on arguments exploiting the geometry underlying preconditioned inverse iteration.     

A subspace preconditioning algorithm for eigenvector/eigenvalue computation

We consider the problem of computing a modest number of the smallest eigenvalues along with orthogonal bases for the corresponding eigenspaces of a symmetric positive definite operatorA defined on a finite dimensional real Hilbert spaceV. In our applications, the dimension ofV is large and the cost of invertingA is prohibitive. In this paper, we shall develop an effective parallelizable technique for computing these eigenvalues and eigenvectors utilizing subspace iteration and preconditioning forA. Estimates will be provided which show that the preconditioned method converges linearly when used with a uniform preconditioner under the assumption that the approximating subspace is close enough to the span of desired eigenvectors.     

The Modified Rayleigh Quotient Iteration

The Rayleigh Quotient Iteration (RQI) is a very popular method for computing eigenpairs of symmetric matrices. It is a special kind of inverse iteration method using the Rayleigh Quotient as shifts. Unfortunately, poor initial approximations may render RQI to slow convergence or even to divergence, In this paper we suggest two kinds of numbers each of which can be used instead of the Rayleigh Quotient as a shifts in the RQI. We call the iteration using the new shifts the Modified Rayleigh Quotient Iteration (MRQI). It has been proved that the MRQI always converges and its convergence rate is cubic.     

Estimating a largest eigenvector by Lanczos and polynomial algorithms with a random start

We study Lanczos and polynomial algorithms with random start for estimating an eigenvector corresponding to the largest eigenvalue of an n × n large symmetric positive definite matrix. We analyze the two error criteria: the randomized error and the randomized residual error. For the randomized error, we prove that it is not possible to get distribution-free bounds, i.e., the bounds must depend on the distribution of eigenvalues of the matrix. We supply such bounds and show that they depend on the ratio of the two largest eigenvalues. For the randomized residual error, distribution-free bounds exist and are provided in the paper. We also provide asymptotic bounds, as well as bounds depending on the ratio of the two largest eigenvalues. The bounds for the Lanczos algorithm may be helpful in a practical implementation and termination of this algorithm. © 1998 John Wiley & Sons, Ltd.     

A symmetry exploiting Lanczos method for symmetric Toeplitz matrices

Several methods for computing the smallest eigenvalues of a symmetric and positive definite Toeplitz matrix T have been studied in the literature. Most of them share the disadvantage that they do not reflect symmetry properties of the corresponding eigenvector. In this note we present a Lanczos method which approximates simultaneously the odd and the even spectrum of T at the same cost as the classical Lanczos approach.     

The relation between the Jacobi algorithm and inverse iteration and a Jacobi algorithm based on elementary reflections

It is shown that the cyclic Jacobi algorithm for the computation of eigenvalues of a symmetric matrix behaves asymptotically like inverse iteration with Rayleigh Quotient shift (RQI). The asymptotic expression for the transformation matrix is used to develop a new Jacobi algorithm which uses elementary reflections (Householder transformations) instead of rotations. The new algorithm has the same asymptotic behaviour, but each sweep needs half the number of arithmetic operations and has one level of looping less than the traditional one. Numerical tests of an APL implementation are reported.     

计算张量特征对的瑞利商迭代法

In this paper,we propose a Rayleigh quotient iteration method (RQI)to calculate the Z-eigenpairs of the symmetric tensor,which can be viewed as a generalization of shifted symmetric higher-order power method (SS-HOPM).The convergence analysis and the fixed-point analysis of this algorithm are given.Nu-merical examples show that RQI needs fewer iterations than SS-HOPM while keep the amount of computation per iteration.     

Computation of eigenvalues and eigenvectors of a symmetric quindiagonal matrix

A recursive algorithm for the implicit derivation of the determinant of a symmetric quindiagonal matrix is developed in terms of its leading principal minors. The algorithm is shown to yield a Sturmian sequence of polynomials from which the eigenvalues can be obtained by use of the bisection process. Further modifications to the inverse iteration method using Wilkinson's technique (1962) yields the required eigenvectors.     

A numerical solution of the constrained weighted energy problem

A numerical algorithm is presented to solve the constrained weighted energy problem from potential theory. As one of the possible applications of this algorithm, we study the convergence properties of the rational Lanczos iteration method for the symmetric eigenvalue problem. The constrained weighted energy problem characterizes the region containing those eigenvalues that are well approximated by the Ritz values. The region depends on the distribution of the eigenvalues, on the distribution of the poles, and on the ratio between the size of the matrix and the number of iterations. Our algorithm gives the possibility of finding the boundary of this region in an effective way.We give numerical examples for different distributions of poles and eigenvalues and compare the results of our algorithm with the convergence behavior of the explicitly performed rational Lanczos algorithm.     

Additive preconditioning, eigenspaces, and the inverse iteration

We incorporate our recent preconditioning techniques into the classical inverse power (Rayleigh quotient) iteration for computing matrix eigenvectors. Every loop of this iteration essentially amounts to solving an ill conditioned linear system of equations. Due to our modification we solve a well conditioned linear system instead. We prove that this modification preserves local quadratic convergence, show experimentally that fast global convergence is preserved as well, and yield similar results for higher order inverse iteration, covering the cases of multiple and clustered eigenvalues.     

A restarted Krylov method with inexact inversions

In this paper, we present a new type of restarted Krylov method for calculating the smallest eigenvalues of a symmetric positive definite matrix G. The new framework avoids the Lanczos tridiagonalization process and the use of polynomial filtering. This simplifies the restarting mechanism and allows the introduction of several modifications. Convergence is assured by a monotonicity property that pushes the eigenvalues toward their limits. Another innovation is the use of inexact inversions of G to generate the Krylov matrices. In this approach, the inverse of G is approximated by using an iterative method to solve the related linear system. Numerical experiments illustrate the usefulness of the proposed approach.     

A reflection on the implicitly restarted Arnoldi method for computing eigenvalues near a vertical line

In this article, we will study the link between a method for computing eigenvalues closest to the imaginary axis and the implicitly restarted Arnoldi method. The extension to computing eigenvalues closest to a vertical line is straightforward, by incorporating a shift. Without loss of generality we will restrict ourselves here to computing eigenvalues closest to the imaginary axis.In a recent publication, Meerbergen and Spence discussed a new approach for detecting purely imaginary eigenvalues corresponding to Hopf bifurcations, which is of interest for the stability of dynamical systems. The novel method is based on inverse iteration (inverse power method) applied on a Lyapunov-like eigenvalue problem. To reduce the computational overhead significantly a projection was added.This method can also be used for computing eigenvalues of a matrix pencil near a vertical line in the complex plane. We will prove in this paper that the combination of inverse iteration with the projection step is equivalent to Sorensen’s implicitly restarted Arnoldi method utilizing well-chosen shifts.     

Extending the eigCG algorithm to nonsymmetric Lanczos for linear systems with multiple right‐hand sides

The technique that was used to build the eigCG algorithm for sparse symmetric linear systems is extended to the nonsymmetric case using the BiCG algorithm. We show that, similar to the symmetric case, we can build an algorithm that is capable of computing a few smallest magnitude eigenvalues and their corresponding left and right eigenvectors of a nonsymmetric matrix using only a small window of the BiCG residuals while simultaneously solving a linear system with that matrix. For a system with multiple right‐hand sides, we give an algorithm that computes incrementally more eigenvalues while solving the first few systems and then uses the computed eigenvectors to deflate BiCGStab for the remaining systems. Our experiments on various test problems, including Lattice QCD, show the remarkable ability of eigBiCG to compute spectral approximations with accuracy comparable with that of the unrestarted, nonsymmetric Lanczos. Furthermore, our incremental eigBiCG followed by appropriately restarted and deflated BiCGStab provides a competitive method for systems with multiple right‐hand sides. Copyright © 2013 John Wiley & Sons, Ltd.     

隐式重启精化全局Lanczos方法

The numerical methods for solving large symmetric eigenvalue problems are considered in this paper.Based on the global Lanczos process,a global Lanczos method for solving large symmetric eigenvalue problems is presented.In order to accelerate the convergence of the F-Ritz vectors,the refined global Lanczos method is developed.Combining the implicitly restarted strategy with the deflation technique,an implicitly restarted and refined global Lanczos method for computing some eigenvalues of large symmetric matrices is proposed.Numerical results show that the proposed methods are efficient.     

Accuracy and effectiveness of the Lanczos algorithm for the symmetric eigenproblem

Eigenvalues and eigenvectors of a large sparse symmetric matrix A can be found accurately and often very quickly using the Lanczos algorithm without reorthogonalization. The algorithm gives essentially correct information on the eigensystem of A, although it does not necessarily give the correct multiplicity of multiple, or even single, eigenvalues. It is straightforward to determine a useful bound on the accuracy of every eigenvalue given by the algorithm. The initial behavior of the algorithm is surprisingly good: it produces vectors spanning the Krylov subspace of a matrix very close to A until this subspace contains an exact eigenvector of a matrix very close to A, and up to this point the effective behavior of the algorithm for the eigenproblem is very like that of the Lanczos algorithm using full reorthogonalization. This helps to explain the remarkable behavior of the basic Lanczos algorithm.     

对称双边对角矩阵特征值问题的计算

1 引 言 大型稀疏矩阵在工程上有广泛的应用.例如,结构工程的有限元分析、电力系统的分析、流体力学及图像数据压缩等应用中常遇到求大型稀疏矩阵的特征值问题.因而矩阵特征值计算问题成为数值代数领域长期关注的问题,如[6][7].最近M.Gu与S.C.Eisenstat     

An implicit QR algorithm for symmetric semiseparable matrices

The QR algorithm is one of the classical methods to compute the eigendecomposition of a matrix. If it is applied on a dense n × n matrix, this algorithm requires O(n³) operations per iteration step. To reduce this complexity for a symmetric matrix to O(n), the original matrix is first reduced to tridiagonal form using orthogonal similarity transformations. In the report (Report TW360, May 2003) a reduction from a symmetric matrix into a similar semiseparable one is described. In this paper a QR algorithm to compute the eigenvalues of semiseparable matrices is designed where each iteration step requires O(n) operations. Hence, combined with the reduction to semiseparable form, the eigenvalues of symmetric matrices can be computed via intermediate semiseparable matrices, instead of tridiagonal ones. The eigenvectors of the intermediate semiseparable matrix will be computed by applying inverse iteration to this matrix. This will be achieved by using an O(n) system solver, for semiseparable matrices. A combination of the previous steps leads to an algorithm for computing the eigenvalue decompositions of semiseparable matrices. Combined with the reduction of a symmetric matrix towards semiseparable form, this algorithm can also be used to calculate the eigenvalue decomposition of symmetric matrices. The presented algorithm has the same order of complexity as the tridiagonal approach, but has larger lower order terms. Numerical experiments illustrate the complexity and the numerical accuracy of the proposed method. Copyright © 2005 John Wiley & Sons, Ltd.     

Application of a block modified Chebyshev algorithm to the iterative solution of symmetric linear systems with multiple right hand side vectors

Summary. An adaptive Richardson iteration method is described for the solution of large sparse symmetric positive definite linear systems of equations with multiple right-hand side vectors. This scheme ``learns' about the linear system to be solved by computing inner products of residual matrices during the iterations. These inner products are interpreted as block modified moments. A block version of the modified Chebyshev algorithm is presented which yields a block tridiagonal matrix from the block modified moments and the recursion coefficients of the residual polynomials. The eigenvalues of this block tridiagonal matrix define an interval, which determines the choice of relaxation parameters for Richardson iteration. Only minor modifications are necessary in order to obtain a scheme for the solution of symmetric indefinite linear systems with multiple right-hand side vectors. We outline the changes required. Received April 22, 1993