DOCSLIB.ORG

A History of Inverse Iteration

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
A History of Inverse Iteration A HISTORY OF INVERSE ITERATION y ILSE C F IPSEN The Metho d Inverse iteration is a metho d for computing eigenfunctions of linear op erators Here we consider inverse iteration for the solution of the algebraic eigenvalue problem Ax x where A is a real or complex square matrix the scalar is an eigenvalue and the non zero vector x is an eigenvector Inverse iteration is the metho d of choice when one has at ones disp osal approximations to a sp ecied subset of eigenvalues of A for which one wants to compute asso ciated eigenvectors The metho d is frequently used in structural mechanics for instance when the extreme eigenvalues and corresp onding eigenvectors of Hermitian p ositivesemidenite eigenproblems are sought Given an approximation to an eigenvalue of A inverse iteration generates a 1 sequence of vectors x from a given starting vector x by solving the linear systems k 0 A I x s x k k k k 1 The scalar s is resp onsible for normalising x Usually one choses s so that kx k k k k k in some norm If everything go es well the sequence of iterates x converges to k an eigenvector asso ciated with an eigenvalue closest to In particular when there is only a single eigenvalue closest to then the x converge to an eigenvector k asso ciated with provided the starting vector x contains a contribution of this 0 k eigenvector Since x is a multiple of the vector A I y and since k 0 1 is the largest eigenvalue of A I the contribution in x of the eigendirection k asso ciated with increases faster as k than the contributions asso ciated with the 2 other eigenvalues Inverse iteration is mathematically identical to the p ower metho d 1 applied to A I In Wilkinson remarked Inverse iteration is now the most widely used metho d for computing eigenvectors corresp onding to selected eigenvalues which have already b een computed more or less accurately p A lo ok at software in the public domain shows that this is still true to day Yet after so many years the b ehaviour of the metho d is still not completely understo o d neither in exact nor in oating p oint arithmetic As a consequence even highly sophisticated state of the art software like LAPACK still comes across matrices for which it cannot compute eigenvectors to high accuracy without a drastic sacrice in eciency 3 This article describ es the eorts undertaken on b ehalf of inverse iteration from its To app ear in Helmut Wielandt Mathematische Werke Mathematical Works Volume II Ma trix Theory and Analysis Hupp ert B and Schneider H eds Walter de Gruyter Berlin y Department of Mathematics North Carolina State University P O Box Raleigh NC USA ipsenmathncsuedu This work was supp orted in part by NSF grant CCR 1 Here I denotes the identity matrix 2 Mathematical identity means that equality holds in exact arithmetic but not necessarily in nite precision arithmetic b ecause the two metho ds determine their iterates by means of dierent op erations 3 We concentrate on the solution of the ordinary eigenvalue problem rather than the generalised eigenvalue problem Ax B x and we do not discuss techniques for convergence acceleration such as Rayleigh quotient iteration where the shift changes with each iteration inception to the present day Octob er Here are a few of the issues that prompted these eorts The iterates may not converge to an eigenvector b ecause the starting vector x do es not have a contribution in the desired eigenvector or b ecause there 0 may b e several dierent eigenvalues of A at the same distance from The eigenvalue closest to may b e a multiple eigenvalue and inverse iteration may have to compute a basis for the corresp onding invariant subspace Unless linear indep endence is enforced explicitly inverse iteration usually fails to compute the desired basis When the matrix is Hermitian linear indep endence is enforced by orthogonalising an iterate against the already computed eigenvectors But dep ending on the shift and the starting vector x iterates may still fail to converge The problem b ecomes even more pronounced 0 when happ ens to approximate a cluster of close but not identical eigenvalues Should one treat the cluster like a collection of separate eigenvalues or like a single multiple eigenvalue or like several multiple eigenvalues How close together must eigenvalues b e to qualify as a cluster Should an iterate b e orthogonalised against all eigenvectors in the cluster or against just a few Wielandt Started It Inverse iteration was intro duced by Wielandt in Although Peters and Wilkinson remark p without further elab oration that a numb er of p eople seem to have had the idea indep endently Wielandt is usually the one credited with the intro duction of inverse iteration He refers to inverse iteration as fractional iteration gebro chene Iteration in German 1 b ecause the matrix A I is a fractional linear function of A p He p oints out the b enet of inverse iteration in the stability analysis of vibrating systems that are small p erturbations of systems whose b ehaviour is known xI In this case go o d approximations to the eigenvalues of the p erturb ed system are available The computational asp ects of Wielandts work leading to the development of inverse iteration are describ ed in the companion pap er Adapted to the algebraic eigenvalue problem Wielandts justication for inverse iteration lo oks as follows xI I Ic Cho ose so that Then Ax x b ecomes A I x x Replacing x on the left by x and on the right by x and writing s instead of k k 1 k gives inverse iteration In Crandall presents inverse iteration for real symmetric matrices as an 4 instance of a particular relaxation metho d He do es not reference Wielandts pap er Wilkinsons Idea The p erson most closely asso ciated with inverse iteration for the algebraic eigenvalue problem is Jim Wilkinson Although Wilkinson gives the credit for inverse iteration to Wielandt he himself came up with it indep endently in as a result of trying to improve Givenss metho d for computing a single eigenvector of a symmetric tridiagonal matrix We mo dify Wilkinsons idea slightly and present it for a general matrix A rather than just a symmetric tridiagonal matrix Supp ose A has order n and the shift is 4 Crandall page gives the following characterisation of relaxation metho ds A unique feature of relaxation metho ds is that they are approximate pro cedures which are not rigidly prescrib ed in advance The precise pro cedure to b e followed is not dictated but is left to the intuition and accumulated exp erience of the computer The computers intelligence is thus an active or dynamic link in the computational chain It is this fact which has made relaxation so attractive to many computers Thus at the time Crandall wrote his pap er a computer was still a human b eing and for exploratory purp oses the relaxation metho ds were carried out with a slide rule p closer to the simple eigenvalue than to any other eigenvalue of A Wilkinsons idea is the following Since an eigenvector x is determined up to a multiple by n of the n equations A I x approximate x by a nonzero vector z that solves n equations of A I z Why should such a z b e a go o d approximation to x Partition A so as to distin guish its leading principal submatrix A of order n and partition z conformally A a z 1 A z a 2 Supp ose is also not an eigenvalue of A so A I is nonsingular If we x to a nonzero value then there exists a solution z to the smaller system A I z a 1 5 This implies that z is a nonzero solution to the larger system A I z e where a z n 2 If e contains a contribution of an eigenvector x then this contribution is amplied n in z by Since is the single closest eigenvalue to the contributions of all other eigenvectors in z are amplied by smaller amounts Thus z is closer to x than is e n Instead of solving the rst n equations one could solve any set of n equations If the ith equation is the one that is omitted then the righthand side is a multiple of e In general the righthand side can b e any vector x as long as it contains a i 0 contribution of a desired eigenvector x Then the solution x of A I x x is 1 1 0 closer to x than is x Hence as inverse iteration progresses the iterates x converge 0 k to an eigenvector x provided the asso ciated eigenvalue is simple Wilkinson Takes Charge The foremost issues on Wilkinsons mind were the choice of appropriate starting vectors and the misfortune of having to solve a linear system that is highly illconditioned when is an accurate approximation to an eigenvalue of A From to Wilkinson analysed the computation of a single eigenvector of a symmetric tridiagonal matrix determination of the convergence rate in exact arithmetic choice of a go o d starting vector in exact arithmetic and eect of the backward error due to the solution of the linear system by Gaussian elimination with partial pivoting on the size of the iterate and the convergence rate pp pp Regarding the computation of eigenvectors asso ciated with almost coincident eigenvalues he recommends a dierent p erturbation of the shift for each eigenvalue and the use of the GramSchmidt algorithm to orthogonalise the current iterate against previously computed eigenvectors p Wilkinson applies a similar treatment to computing a single eigenvector when the eigenvalue is semisimple and the matrix is real and nonsymmetric pp As Wielandt had already observed in the context of the p ower metho
Load more

Recommended publications
  • A Geometric Theory for Preconditioned Inverse Iteration. III: a Short and Sharp Convergence Estimate for Generalized Eigenvalue Problems
    A geometric theory for preconditioned inverse iteration. III: A short and sharp convergence estimate for generalized eigenvalue problems. Andrew V. Knyazev Department of Mathematics, University of Colorado at Denver, P.O. Box 173364, Campus Box 170, Denver, CO 80217-3364 1 Klaus Neymeyr Mathematisches Institut der Universit¨atT¨ubingen,Auf der Morgenstelle 10, 72076 T¨ubingen,Germany 2 Abstract In two previous papers by Neymeyr: A geometric theory for preconditioned inverse iteration I: Extrema of the Rayleigh quotient, LAA 322: (1-3), 61-85, 2001, and A geometric theory for preconditioned inverse iteration II: Convergence estimates, LAA 322: (1-3), 87-104, 2001, a sharp, but cumbersome, convergence rate estimate was proved for a simple preconditioned eigensolver, which computes the smallest eigenvalue together with the corresponding eigenvector of a symmetric positive def- inite matrix, using a preconditioned gradient minimization of the Rayleigh quotient. In the present paper, we discover and prove a much shorter and more elegant, but still sharp in decisive quantities, convergence rate estimate of the same method that also holds for a generalized symmetric definite eigenvalue problem. The new estimate is simple enough to stimulate a search for a more straightforward proof technique that could be helpful to investigate such practically important method as the locally optimal block preconditioned conjugate gradient eigensolver. We demon- strate practical effectiveness of the latter for a model problem, where it compares favorably with two
    [Show full text]
  • The Godunov–Inverse Iteration: a Fast and Accurate Solution to the Symmetric Tridiagonal Eigenvalue Problem
    The Godunov{Inverse Iteration: A Fast and Accurate Solution to the Symmetric Tridiagonal Eigenvalue Problem Anna M. Matsekh a;1 aInstitute of Computational Technologies, Siberian Branch of the Russian Academy of Sciences, Lavrentiev Ave. 6, Novosibirsk 630090, Russia Abstract We present a new hybrid algorithm based on Godunov's method for computing eigenvectors of symmetric tridiagonal matrices and Inverse Iteration, which we call the Godunov{Inverse Iteration. We use eigenvectors computed according to Go- dunov's method as starting vectors in the Inverse Iteration, replacing any non- numeric elements of Godunov's eigenvectors with random uniform numbers. We use the right-hand bounds of the Ritz intervals found by the bisection method as Inverse Iteration shifts, while staying within guaranteed error bounds. In most test cases convergence is reached after only one step of the iteration, producing error estimates that are as good as or superior to those produced by standard Inverse Iteration implementations. Key words: Symmetric eigenvalue problem, tridiagonal matrices, Inverse Iteration 1 Introduction Construction of algorithms that enable to find all eigenvectors of the symmet- ric tridiagonal eigenvalue problem with guaranteed accuracy in O(n2) arith- metic operations has become one of the most pressing problems of the modern numerical algebra. QR method, one of the most accurate methods for solv- ing eigenvalue problems, requires 6n3 arithmetic operations and O(n2) square root operations to compute all eigenvectors of a tridiagonal matrix [Golub and Email address: [email protected] (Anna M. Matsekh ). 1 Present address: Modeling, Algorithms and Informatics Group, Los Alamos Na- tional Laboratory. P.O. Box 1663, MS B256, Los Alamos, NM 87544, USA Preprint submitted to Elsevier Science 31 January 2003 Loan (1996)].
    [Show full text]
  • A Note on Inverse Iteration
    A Note on Inverse Iteration Klaus Neymeyr Universit¨atRostock, Fachbereich Mathematik, Universit¨atsplatz 1, 18051 Rostock, Germany; SUMMARY Inverse iteration, if applied to a symmetric positive definite matrix, is shown to generate a sequence of iterates with monotonously decreasing Rayleigh quotients. We present sharp bounds from above and from below which highlight inverse iteration as a descent scheme for the Rayleigh quotient. Such estimates provide the background for the analysis of the behavior of the Rayleigh quotient in certain approximate variants of inverse iteration. key words: Symmetric eigenvalue problem; Inverse iteration; Rayleigh quotient. 1. Introduction Inverse iteration is a well-known iterative procedure to compute approximations of eigenfunctions and eigenvalues of linear operators. It was introduced by Wielandt in 1944 in a sequence of five papers, see [1], to treat the matrix eigenvalue problem Axi = λixi for a real or complex square matrix A. The scalar λi is the ith eigenvalue and the vector xi denotes a corresponding eigenvector. Given a nonzero starting vector x(0), inverse iteration generates a sequence of iterates x(k) by solving the linear systems (A − σI)x(k+1) = x(k), k = 0, 1, 2,..., (1) where σ denotes an eigenvalue approximation and I is the identity matrix. In practice, the iterates are normalized after each step. If A is a symmetric matrix, then the iterates x(k) converge to an eigenvector associated with an eigenvalue closest to σ if the starting vector x(0) is not perpendicular to that vector. For non-symmetric matrices the issue of starting vectors is discussed in Sec.
    [Show full text]
  • Inverse, Shifted Inverse, and Rayleigh Quotient Iteration As Newton’S Method⇤
    INVERSE, SHIFTED INVERSE, AND RAYLEIGH QUOTIENT ITERATION AS NEWTON’S METHOD⇤ JOHN DENNIS JR.† AND RICHARD TAPIA‡ Abstract. Two-norm normalized inverse, shifted inverse, and Rayleigh quotient iteration are well-known algorithms for approximating an eigenvector of a symmetric matrix. In this work we establish rigorously that each one of these three algorithms can be viewed as a standard form of Newton’s method from the nonlinear programming literature, followed by the normalization. This equivalence adds considerable understanding to the formal structure of inverse, shifted inverse, and Rayleigh quotient iteration and provides an explanation for their good behavior despite the possible need to solve systems with nearly singular coefficient matrices; the algorithms have what can be viewed as removable singularities. A thorough historical development of these eigenvalue algorithms is presented. Utilizing our equivalences we construct traditional Newton’s method theory analysis in order to gain understanding as to why, as normalized Newton’s method, inverse iteration and shifted inverse iteration are only linearly convergent and not quadratically convergent, and why a new linear system need not be solved at each iteration. We also investigate why Rayleigh quotient iteration is cubically convergent and not just quadratically convergent. Key words. inverse power method, inverse iteration, shifted inverse iteration, Rayleigh quotient iteration, Newton’s method AMS subject classifications. 65F15, 49M37, 49M15, 65K05 1. Introduction. The initial
    [Show full text]
  • Arxiv:1105.1185V1 [Math.NA] 5 May 2011 Ento 2.2
    ITERATIVE METHODS FOR COMPUTING EIGENVALUES AND EIGENVECTORS MAYSUM PANJU Abstract. We examine some numerical iterative methods for computing the eigenvalues and eigenvectors of real matrices. The five methods examined here range from the simple power iteration method to the more complicated QR iteration method. The derivations, procedure, and advantages of each method are briefly discussed. 1. Introduction Eigenvalues and eigenvectors play an important part in the applications of linear algebra. The naive method of finding the eigenvalues of a matrix involves finding the roots of the characteristic polynomial of the matrix. In industrial sized matrices, however, this method is not feasible, and the eigenvalues must be obtained by other means. Fortunately, there exist several other techniques for finding eigenvalues and eigenvectors of a matrix, some of which fall under the realm of iterative methods. These methods work by repeatedly refining approximations to the eigenvectors or eigenvalues, and can be terminated whenever the approximations reach a suitable degree of accuracy. Iterative methods form the basis of much of modern day eigenvalue computation. In this paper, we outline five such iterative methods, and summarize their derivations, procedures, and advantages. The methods to be examined are the power iteration method, the shifted inverse iteration method, the Rayleigh quotient method, the simultaneous iteration method, and the QR method. This paper is meant to be a survey over existing algorithms for the eigenvalue computation problem. Section 2 of this paper provides a brief review of some of the linear algebra background required to understand the concepts that are discussed. In section 3, the iterative methods are each presented, in order of complexity, and are studied in brief detail.
    [Show full text]
  • Preconditioned Inverse Iteration and Shift-Invert Arnoldi Method
    Preconditioned inverse iteration and shift-invert Arnoldi method Melina Freitag Department of Mathematical Sciences University of Bath CSC Seminar Max-Planck-Institute for Dynamics of Complex Technical Systems Magdeburg 3rd May 2011 joint work with Alastair Spence (Bath) 1 Introduction 2 Inexact inverse iteration 3 Inexact Shift-invert Arnoldi method 4 Inexact Shift-invert Arnoldi method with implicit restarts 5 Conclusions Outline 1 Introduction 2 Inexact inverse iteration 3 Inexact Shift-invert Arnoldi method 4 Inexact Shift-invert Arnoldi method with implicit restarts 5 Conclusions Problem and iterative methods Find a small number of eigenvalues and eigenvectors of: Ax = λx, λ ∈ C,x ∈ Cn A is large, sparse, nonsymmetric Problem and iterative methods Find a small number of eigenvalues and eigenvectors of: Ax = λx, λ ∈ C,x ∈ Cn A is large, sparse, nonsymmetric Iterative solves Power method Simultaneous iteration Arnoldi method Jacobi-Davidson method repeated application of the matrix A to a vector Generally convergence to largest/outlying eigenvector Problem and iterative methods Find a small number of eigenvalues and eigenvectors of: Ax = λx, λ ∈ C,x ∈ Cn A is large, sparse, nonsymmetric Iterative solves Power method Simultaneous iteration Arnoldi method Jacobi-Davidson method The first three of these involve repeated application of the matrix A to a vector Generally convergence to largest/outlying eigenvector Shift-invert strategy Wish to find a few eigenvalues close to a shift σ σ λλ λ λ λ 3 1 2 4 n Shift-invert strategy Wish to find a few eigenvalues close to a shift σ σ λλ λ λ λ 3 1 2 4 n Problem becomes − 1 (A − σI) 1x = x λ − σ each step of the iterative method involves repeated application of A =(A − σI)−1 to a vector Inner iterative solve: (A − σI)y = x using Krylov method for linear systems.
    [Show full text]
  • Lecture 12 — February 26 and 28 12.1 Introduction
    EE 381V: Large Scale Learning Spring 2013 Lecture 12 | February 26 and 28 Lecturer: Caramanis & Sanghavi Scribe: Karthikeyan Shanmugam and Natalia Arzeno 12.1 Introduction In this lecture, we focus on algorithms that compute the eigenvalues and eigenvectors of a real symmetric matrix. Particularly, we are interested in finding the largest and smallest eigenvalues and the corresponding eigenvectors. We study two methods: Power method and the Lanczos iteration. The first involves multiplying the symmetric matrix by a randomly chosen vector, and iteratively normalizing and multiplying the matrix by the normalized vector from the previous step. The convergence is geometric, i.e. the `1 distance between the true and the computed largest eigenvalue at the end of every step falls geometrically in the number of iterations and the rate depends on the ratio between the second largest and the largest eigenvalue. Some generalizations of the power method to compute the largest k eigenvalues and the eigenvectors will be discussed. The second method (Lanczos iteration) terminates in n iterations where each iteration in- volves estimating the largest (smallest) eigenvalue by maximizing (minimizing) the Rayleigh coefficient over vectors drawn from a suitable subspace. At each iteration, the dimension of the subspace involved in the optimization increases by 1. The sequence of subspaces used are Krylov subspaces associated with a random initial vector. We study the relation between Krylov subspaces and tri-diagonalization of a real symmetric matrix. Using this connection, we show that an estimate of the extreme eigenvalues can be computed at each iteration which involves eigen-decomposition of a tri-diagonal matrix.
    [Show full text]
  • AMS526: Numerical Analysis I (Numerical Linear Algebra) Lecture 16: Rayleigh Quotient Iteration
    AMS526: Numerical Analysis I (Numerical Linear Algebra) Lecture 16: Rayleigh Quotient Iteration Xiangmin Jiao SUNY Stony Brook Xiangmin Jiao Numerical Analysis I 1 / 10 Solving Eigenvalue Problems All eigenvalue solvers must be iterative Iterative algorithms have multiple facets: 1 Basic idea behind the algorithms 2 Convergence and techniques to speed-up convergence 3 Efficiency of implementation 4 Termination criteria We will focus on first two aspects Xiangmin Jiao Numerical Analysis I 2 / 10 Simplification: Real Symmetric Matrices We will consider eigenvalue problems for real symmetric matrices, i.e. T m×m m A = A 2 R , and Ax = λx for x 2 R p ∗ T T I Note: x = x , and kxk = x x A has real eigenvalues λ1,λ2, ::: , λm and orthonormal eigenvectors q1, q2, ::: , qm, where kqj k = 1 Eigenvalues are often also ordered in a particular way (e.g., ordered from large to small in magnitude) In addition, we focus on symmetric tridiagonal form I Why? Because phase 1 of two-phase algorithm reduces matrix into tridiagonal form Xiangmin Jiao Numerical Analysis I 3 / 10 Rayleigh Quotient m The Rayleigh quotient of x 2 R is the scalar xT Ax r(x) = xT x For an eigenvector x, its Rayleigh quotient is r(x) = xT λx=xT x = λ, the corresponding eigenvalue of x For general x, r(x) = α that minimizes kAx − αxk2. 2 x is eigenvector of A() rr(x) = x T x (Ax − r(x)x) = 0 with x 6= 0 r(x) is smooth and rr(qj ) = 0 for any j, and therefore is quadratically accurate: 2 r(x) − r(qJ ) = O(kx − qJ k ) as x ! qJ for some J Xiangmin Jiao Numerical Analysis I 4 / 10 Power
    [Show full text]
  • Randomized Numerical Linear Algebra: Foundations & Algorithms
    Randomized Numerical Linear Algebra: Foundations & Algorithms Per-Gunnar Martinsson, University of Texas at Austin Joel A. Tropp, California Institute of Technology Abstract: This survey describes probabilistic algorithms for linear algebra computations, such as factorizing matrices and solving linear systems. It focuses on techniques that have a proven track record for real-world problems. The paper treats both the theoretical foundations of the subject and practical computational issues. Topics include norm estimation; matrix approximation by sampling; structured and un- structured random embeddings; linear regression problems; low-rank approximation; sub- space iteration and Krylov methods; error estimation and adaptivity; interpolatory and CUR factorizations; Nystrom¨ approximation of positive semidefinite matrices; single- view (“streaming”) algorithms; full rank-revealing factorizations; solvers for linear sys- tems; and approximation of kernel matrices that arise in machine learning and in scientific computing. CONTENTS 1. Introduction 2 2. Linear algebra preliminaries 6 3. Probability preliminaries 9 4. Trace estimation by sampling 10 5. Schatten p-norm estimation by sampling 16 6. Maximum eigenvalues and trace functions 18 7. Matrix approximation by sampling 24 8. Randomized embeddings 30 9. Structured random embeddings 35 10. How to use random embeddings 40 arXiv:2002.01387v2 [math.NA] 15 Aug 2020 11. The randomized rangefinder 43 12. Error estimation and adaptivity 52 13. Finding natural bases: QR, ID, and CUR 57 14. Nystrom¨ approximation 62 15. Single-view algorithms 64 16. Factoring matrices of full or nearly full rank 68 17. General linear solvers 76 18. Linear solvers for graph Laplacians 79 19. Kernel matrices in machine learning 85 20. High-accuracy approximation of kernel matrices 94 References 106 Randomized Numerical Linear Algebra P.G.
    [Show full text]
  • Computing Selected Eigenvalues of Sparse Unsymmetric Matrices Using Subspace Iteration
    RAL-91-056 Computing selected eigenvalues of sparse unsymmetric matrices using subspace iteration by I. S. Duff and J. A. Scott ABSTRACT This paper discusses the design and development of a code to calculate the eigenvalues of a large sparse real unsymmetric matrix that are the right-most, left-most, or are of largest modulus. A subspace iteration algorithm is used to compute a sequence of sets of vectors that converge to an orthonormal basis for the invariant subspace corresponding to the required eigenvalues. This algorithm is combined with Chebychev acceleration if the right-most or left-most eigenvalues are sought, or if the eigenvalues of largest modulus are known to be the right-most or left-most eigenvalues. An option exists for computing the corresponding eigenvectors. The code does not need the matrix explicitly since it only requires the user to multiply sets of vectors by the matrix. Sophisticated and novel iteration controls, stopping criteria, and restart facilities are provided. The code is shown to be efficient and competitive on a range of test problems. Central Computing Department, Atlas Centre, Rutherford Appleton Laboratory, Oxon OX11 0QX. August 1993. 1 Introduction We are concerned with the problem of computing selected eigenvalues and the corresponding eigenvectors of a large sparse real unsymmetric matrix. In particular, we are interested in computing either the eigenvalues of largest modulus or the right-most (or left-most) eigenvalues. This problem arises in a significant number of applications, including mathematical models in economics, Markov chain modelling of queueing networks, and bifurcation problems (for references, see Saad 1984).
    [Show full text]
  • Arxiv:1803.03717V1 [Math.NA] 9 Mar 2018 Pression Techniques, the Iterates Are Forced to Stay in Low-Rank Format
    LOW-RANK SOLUTION METHODS FOR STOCHASTIC EIGENVALUE PROBLEMS∗ HOWARD C. ELMANy AND TENGFEI SUz Abstract. We study efficient solution methods for stochastic eigenvalue problems arising from discretization of self-adjoint partial differential equations with random data. With the stochastic Galerkin approach, the solutions are represented as generalized polynomial chaos expansions. A low-rank variant of the inverse subspace iteration algorithm is presented for computing one or sev- eral minimal eigenvalues and corresponding eigenvectors of parameter-dependent matrices. In the algorithm, the iterates are approximated by low-rank matrices, which leads to significant cost sav- ings. The algorithm is tested on two benchmark problems, a stochastic diffusion problem with some poorly separated eigenvalues, and an operator derived from a discrete stochastic Stokes problem whose minimal eigenvalue is related to the inf-sup stability constant. Numerical experiments show that the low-rank algorithm produces accurate solutions compared to the Monte Carlo method, and it uses much less computational time than the original algorithm without low-rank approximation. Key words. stochastic eigenvalue problem, inverse subspace iteration, low-rank approximation AMS subject classifications. 35R60, 65F15, 65F18, 65N22 1. Introduction. Approaches for solving stochastic eigenvalue problems can be broadly divided into non-intrusive methods, including Monte Carlo methods and stochastic collocation methods [1, 30], and intrusive stochastic Galerkin methods. The Galerkin approach gives parametrized descriptions of the eigenvalues and eigen- vectors, represented as expansions with stochastic basis functions. A commonly used framework is the generalized polynomial chaos (gPC) expansion [40]. A direct projec- tion onto the basis functions will result in large coupled nonlinear systems that can be solved by a Newton-type algorithm [6, 13].
    [Show full text]
  • AM 205: Lecture 22
    AM 205: lecture 22 I Final project proposal due by 6pm on Thu Nov 17. Email Chris or the TFs to set up a meeting. Those who have completed this will see four proposal points on Canvas. I Today: eigenvalue algorithms, QR algorithm Sensitivity of Eigenvalue Problems Weyl's Theorem: Let λ1 ≤ λ2 ≤ · · · ≤ λn and λ~1 ≤ λ~2 ≤ · · · ≤ λ~n be the eigenvalues of hermitian matrices A and A + δA, respectively. Then max jλi − λ~i j ≤ kδAk2. i=1;:::;n Hence in the hermitian case, each perturbed eigenvalue must be in the disk1 of its corresponding unperturbed eigenvalue! 1In fact, eigenvalues of a hermitian matrix are real, so disk here is actually an interval in R Sensitivity of Eigenvalue Problems The Bauer{Fike Theorem relates to perturbations of the whole spectrum We can also consider perturbations of individual eigenvalues n×n Suppose, for simplicity, that A 2 C is symmetric, and consider the perturbed eigenvalue problem (A + E)(v + ∆v) = (λ + ∆λ)(v + ∆v) Expanding this equation, dropping second order terms, and using Av = λv gives A∆v + Ev ≈ ∆λv + λ∆v Sensitivity of Eigenvalue Problems Premultiply A∆v + Ev ≈ ∆λv + λ∆v by v ∗ to obtain v ∗A∆v + v ∗Ev ≈ ∆λv ∗v + λv ∗∆v Noting that v ∗A∆v = (v ∗A∆v)∗ = ∆v ∗Av = λ∆v ∗v = λv ∗∆v leads to v ∗Ev v ∗Ev ≈ ∆λv ∗v; or∆ λ = v ∗v Sensitivity of Eigenvalue Problems Finally, we obtain ∗ jv Evj kvk2kEvk2 j∆λj ≈ 2 ≤ 2 = kEk2; kvk2 kvk2 so that j∆λj . kEk2 We observe that I perturbation bound does not depend on cond(V ) when we consider only an individual eigenvalue I this individual eigenvalue perturbation bound
    [Show full text]