双曲矩阵分解修正问题的一些结果

This paper considers the updating problem of the hyperbolic matrix factorizations. The sufficient conditions for the existence of the updated hyperbolic matrix factorizations are first provided. Then, some differential inequalities and first order perturbation expansions for the updated hyperbolic factors are derived. These results generalize the corresponding ones for the updating problem of the classical QR factorization obtained by Jiguang SUN.     

The Equality Constrained Indefinite Least Squares Problem: Theory and Algorithms

We present theory and algorithms for the equality constrained indefinite least squares problem, which requires minimization of an indefinite quadratic form subject to a linear equality constraint. A generalized hyperbolic QR factorization is introduced and used in the derivation of perturbation bounds and to construct a numerical method. An alternative method is obtained by employing a generalized QR factorization in combination with a Cholesky factorization. Rounding error analysis is given to show that both methods have satisfactory numerical stability properties and numerical experiments are given for illustration. This work builds on recent work on the unconstrained indefinite least squares problem by Chandrasekaran, Gu, and Sayed and by the present authors.     

On the modified Gram-Schmidt algorithm for weighted and constrained linear least squares problems

A framework and an algorithm for using modified Gram-Schmidt for constrained and weighted linear least squares problems is presented. It is shown that a direct implementation of a weighted modified Gram-Schmidt algorithm is unstable for heavily weighted problems. It is shown that, in most cases it is possible to get a stable algorithm by a simple modification free from any extra computational costs. In particular, it is not necessary to perform reorthogonalization.Solving the weighted and constrained linear least squares problem with the presented weighted modified Gram-Schmidt algorithm is seen to be numerically equivalent to an algorithm based on a weighted Householder-likeQR factorization applied to a slightly larger problem. This equivalence is used to explain the instability of the weighted modified Gram-Schmidt algorithm. If orthogonality, with respect to a weighted inner product, of the columns inQ is important then reorthogonalization can be used. One way of performing such reorthogonalization is described.Computational tests are given to show the main features of the algorithm.     

A generalization of Bartlett’s decomposition

Bartlett’s decomposition provides the distributional properties of the elements of the Cholesky factor of A=G^TG where the elements of G are i.i.d. standard Gaussian random variables. In this paper the most general case where the elements of G have a joint multivariate Gaussian density is considered.     

Multifrontal QR Factorization in a Multiprocessor Environment

We describe the design and implementation of a parallel QR decomposition algorithm for a large sparse matrix A . The algorithm is based on the multifrontal approach and makes use of Householder transformations. The tasks are distributed among processors according to an assembly tree which is built from the symbolic factorization of the matrix A ^T A . We first address uniprocessor issues and then discuss the multiprocessor implementation of the method. We consider the parallelization of both the factorization phase and the solve phase. We use relaxation of the sparsity structure of both the original matrix and the frontal matrices to improve the performance. We show that, in this case, the use of Level 3 BLAS can lead to very significant gains in performance. We use the eight processor Alliant˜FX/80 at CERFACS to illustrate our discussion.     

Bounds on Singular Values Revealed by QR Factorizations

We introduce a pair of dual concepts: pivoted blocks and reverse pivoted blocks. These blocks are the outcome of a special column pivoting strategy in QR factorization. Our main result is that under such a column pivoting strategy, the QR factorization of a given matrix can give tight estimates of any two a priori-chosen consecutive singular values of that matrix. In particular, a rank-revealing QR factorization is guaranteed when the two chosen consecutive singular values straddle a gap in the singular value spectrum that gives rise to the rank degeneracy of the given matrix. The pivoting strategy, called cyclic pivoting, can be viewed as a generalization of Golub's column pivoting and Stewart's reverse column pivoting. Numerical experiments confirm the tight estimates that our theory asserts.     

Regularization Using QR Factorization and the Estimation of the Optimal Parameter

In this paper we propose a direct regularization method using QR factorization for solving linear discrete ill-posed problems. The decomposition of the coefficient matrix requires less computational cost than the singular value decomposition which is usually used for Tikhonov regularization. This method requires a parameter which is similar to the regularization parameter of Tikhonov's method. In order to estimate the optimal parameter, we apply three well-known parameter choice methods for Tikhonov regularization.This revised version was published online in October 2005 with corrections to the Cover Date.     

A note on the recursive calculation of dominant singular subspaces

A recursive procedure for computing an approximation of the left and right dominant singular subspaces of a given matrix is proposed in [1]. The method is particularly suited for matrices with many more rows than columns. The procedure consists of a few steps. In one of these steps a Householder transformation is multiplied to an upper triangular matrix. The following step consists in recomputing an upper triangular matrix from the latter product. In [1] it is said that the latter step is accomplished in O(k³) operations, where k is the order of the triangular matrix. In this short note we show that this step can be accomplished in O(k²) operations. This research was partially supported by MIUR, grant number 2002014121 (first author) and by the Research Council K.U.Leuven, project OT/00/16 (SLAP: Structured Linear Algebra Package), by the Fund for Scientific Research–Flanders (Belgium), projects G.0078.01 (SMA: Structured Matrices and their Applications), G.0176.02 (ANCILA: Asymptotic aNalysis of the Convergence behavior of Iterative methods in numerical Linear Algebra), G.0184.02 (CORFU: Constructive study of Orthogonal Functions) and G.0455.0 (RHPH: Riemann–Hilbert problems, random matrices and Padé–Hermite approximation), and by the Belgian Programme on Interuniversity Poles of Attraction, initiated by the Belgian State, Prime Ministers Office for Science, Technology and Culture, project IUAP V-22 (Dynamical Systems and Control: Computation, Identification & Modelling) (second and third author). The scientific responsibility rests with the authors.AMS subject classification 15A15, 15A09, 15A23     

An Exact Least Trimmed Squares Algorithm for a Range of Coverage Values

A new algorithm to solve exact least trimmed squares (LTS) regression is presented. The adding row algorithm (ARA) extends existing methods that compute the LTS estimator for a given coverage. It employs a tree-based strategy to compute a set of LTS regressors for a range of coverage values. Thus, prior knowledge of the optimal coverage is not required. New nodes in the regression tree are generated by updating the QR decomposition of the data matrix after adding one observation to the regression model. The ARA is enhanced by employing a branch and bound strategy. The branch and bound algorithm is an exhaustive algorithm that uses a cutting test to prune nonoptimal subtrees. It significantly improves over the ARA in computational performance. Observation preordering throughout the traversal of the regression tree is investigated. A computationally efficient and numerically stable calculation of the bounds using Givens rotations is designed around the QR decomposition, avoiding the need to explicitly update the triangular factor when an observation is added. This reduces the overall computational load of the preordering device by approximately half. A solution is proposed to allow preordering when the model is underdetermined. It employs pseudo-orthogonal rotations to downdate the QR decomposition. The strategies are illustrated by example. Experimental results confirm the computational efficiency of the proposed algorithms. Supplemental materials (R package and formal proofs) are available online.     

A Lanczos‐type algorithm for the QR factorization of regular Cauchy matrices

We present a fast algorithm for computing the QR factorization of Cauchy matrices with real nodes. The algorithm works for almost any input matrix, does not require squaring the matrix, and fully exploits the displacement structure of Cauchy matrices. We prove that, if the determinant of a certain semiseparable matrix is non‐zero, a three term recurrence relation among the rows or columns of the factors exists. Copyright © 2002 John Wiley & Sons, Ltd.     

基于反Krylov矩阵正交分解的半可分矩阵

在本文中,我们证明了对一个反Krylov矩阵作QR分解后,利用得到的正交矩阵可以将一个具有互异特征值的对称矩阵转化为一个半可分矩阵的形式,这个结果表明了反Krylov矩阵与半可分矩阵之间的联系．另外,我们还证明了这类对称半可分矩阵在QR达代下矩阵结构保持不变性．     

SR分解的乘法扰动界

In this paper, we present the first order perturbation bounds for the SR factorization with respect to left multiplicative perturbation, and the first order and rigorous perturbation bounds for this factorization with respect to right multiplicative perturbation.Moreover, taking the properties of SR factors into consideration, we also provide some refined perturbation bounds.     

Stability analysis of a general toeplitz systems solver

We show that a fast algorithm for theQR factorization of a Toeplitz or Hankel matrixA is weakly stable in the sense thatR ^T R is close toA ^T A. Thus, when the algorithm is used to solve the semi-normal equationsR ^TRx=A^T b, we obtain a weakly stable method for the solution of a nonsingular Toeplitz or Hankel linear systemAx=b. The algorithm also applies to the solution of the full-rank Toeplitz or Hankel least squares problem min ||Ax-b||₂.     

Continuity of the null space basis and constrained optimization

Many constrained optimization algorithms use a basis for the null space of the matrix of constraint gradients. Recently, methods have been proposed that enable this null space basis to vary continuously as a function of the iterates in a neighborhood of the solution. This paper reports results from topology showing that, in general, there is no continuous function that generates the null space basis of all full rank rectangular matrices of a fixed size. Thus constrained optimization algorithms cannot assume an everywhere continuous null space basis. We also give some indication of where these discontinuities must occur. We then propose an alternative implementation of a class of constrained optimization algorithms that uses approximations to the reduced Hessian of the Lagrangian but is independent of the choice of null space basis. This approach obviates the need for a continuously varying null space basis.Research supported by NSF grant MCS 81-15475 and DCR-8403483Research supported by ARO contracts DAAG 29-81-K-0108 and DAAG 29-84-K-0140     

Sparse Linear Least Squares Problems in Optimization

Numerical and computational aspects of direct methods for largeand sparseleast squares problems are considered. After a brief survey of the most oftenused methods, we summarize the important conclusions made from anumerical comparison in matlab. Significantly improved algorithms haveduring the last 10-15 years made sparse QR factorization attractive, andcompetitive to previously recommended alternatives. Of particular importanceis the multifrontal approach, characterized by low fill-in, dense subproblemsand naturally implemented parallelism. We describe a Householder multifrontalscheme and its implementation on sequential and parallel computers. Availablesoftware has in practice a great influence on the choice of numericalalgorithms. Less appropriate algorithms are thus often used solely because ofexisting software packages. We briefly survey softwarepackages for the solution of sparse linear least squares problems. Finally,we focus on various applications from optimization, leading to the solution oflarge and sparse linear least squares problems. In particular, we concentrateon the important case where the coefficient matrix is a fixed general sparsematrix with a variable diagonal matrix below. Inner point methods forconstrained linear least squares problems give, for example, rise to suchsubproblems. Important gains can be made by taking advantage of structure.Closely related is also the choice of numerical method for these subproblems.We discuss why the less accurate normal equations tend to be sufficient inmany applications.     

QR factorization for row or column symmetric matrix

The problem of fast computing the QR factorization of row or column symmetric matrix is considered. We address two new algorithms based on a correspondence of Q and R matrices between the row or column symmetric matrix and its mother matrix. Theoretical analysis and numerical evidence show that, for a class of row or column symmetric matrices, the QR factorization using the mother matrix rather than the row or column symmetric matrix per se can save dramatically the CPU time and memory without loss of any numerical precision.