On tridiagonal matrices unitarily equivalent to normal matrices

In this article the unitary equivalence transformation of normal matrices to tridiagonal form is studied.It is well-known that any matrix is unitarily equivalent to a tridiagonal matrix. In case of a normal matrix the resulting tridiagonal inherits a strong relation between its super- and subdiagonal elements. The corresponding elements of the super- and subdiagonal will have the same absolute value.In this article some basic facts about a unitary equivalence transformation of an arbitrary matrix to tridiagonal form are firstly studied. Both an iterative reduction based on Krylov sequences as a direct tridiagonalization procedure via Householder transformations are reconsidered. This equivalence transformation is then applied to the normal case and equality of the absolute value between the super- and subdiagonals is proved. Self-adjointness of the resulting tridiagonal matrix with regard to a specific scalar product is proved. Properties when applying the reduction on symmetric, skew-symmetric, Hermitian, skew-Hermitian and unitary matrices and their relations with, e.g., complex symmetric and pseudo-symmetric matrices are presented.It is shown that the reduction can then be used to compute the singular value decomposition of normal matrices making use of the Takagi factorization. Finally some extra properties of the reduction as well as an efficient method for computing a unitary complex symmetric decomposition of a normal matrix are given.     

On Parallel Fast Jacobi Algorithm for the Eigenproblem of Real Symmetric Matrices

Jacobi algorithm has been developed for the eigenproblem of real symmetric matrices, singular value decomposition of matrices and least squares of the overdetermined system on a parallel computer. In this paper, the parallel schemes and fast algorithm are discussed, and the error analysis and a new bound are presented.     

A sort-Jacobi algorithm for semisimple lie algebras

A structure preserving sort-Jacobi algorithm for computing eigenvalues or singular values is presented. It applies to an arbitrary semisimple Lie algebra on its (−1)-eigenspace of the Cartan involution. Local quadratic convergence for arbitrary cyclic schemes is shown for the regular case. The proposed method is independent of the representation of the underlying Lie algebra and generalizes well-known normal form problems such as e.g. the symmetric, Hermitian, skew-symmetric, symmetric and skew-symmetric R-Hamiltonian eigenvalue problem and the singular value decomposition.     

Perturbation Theory for Simultaneous Bases of Singular Subspaces

New perturbation theorems are proved for simultaneous bases of singular subspaces of real matrices. These results improve the absolute bounds previously obtained in [6] for general (complex) matrices. Unlike previous results, which are valid only for the Frobenius norm, the new bounds, as well as those in [6] for complex matrices, are extended to any unitarily invariant matrix norm. The bounds are complemented with numerical experiments which show their relevance for the algorithms computing the singular value decomposition. Additionally, the differential calculus approach employed allows to easily prove new sin perturbation theorems for singular subspaces which deal independently with left and right singular subspaces.     

Bidiagonal decomposition of totally positive Bernstein-Vandermonde matrices

The class of Bernstein-Vandermonde matrices (a generalization of Vandermonde matrices arising when the monomial basis is replaced by the Bernstein basis) is considered. A convenient ordering of their rows makes these matrices strictly totally positive. By using results related to total positivity and Neville elimination, an algorithm for computing the bidiagonal decomposition of a Bernstein-Vandermonde matrix is constructed. The use of explicit expressions for the determinants involved in the process serves to make the algorithm both fast and accurate. One of the applications of our algorithm is the design of fast and accurate algorithms for solving Lagrange interpolation problems when using the Bernstein basis, an approach useful for the field of Computer Aided Geometric Design since it avoids the stability problems involved with basis transformations between the Bernstein and the monomial bases. A different application consists of the use of the bidiagonal decomposition as an intermediate step of the computation of the eigenvalues and the singular value decomposition of a totally positive Bernstein-Vandermonde matrix. (© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)     

Efficient diagonalization of oversized matrices on a distributed-memory multiprocessor

On parallel architectures, Jacobi methods for computing the singular value decomposition (SVD) and the symmetric eigenvalue decomposition (EVD) have been established as one of the most popular algorithms due to their excellent parallelism. Most of the Jacobi algorithms for distributed-memory architectures have been developed under the assumption that matrices can be distributed over the processors by square blocks of an even order or column blocks with an even number of columns. Obviously, there is a limit on the number of processors while we need to deal with problems of various sizes. We propose algorithms to diagonalize oversized matrices on a given distributed-memory multiprocessor with good load balancing and minimal message passing. Performances of the proposed algorithms vary greatly, depending on the relation between the problem size and the number of available processors. We give theoretical performance analyses which suggest the faster algorithm for a given problem size on a given distributed-memory multiprocessor. Finally, we present a new implementation for the convergence test of the algorithms on a distributed-memory multiprocessor and the implementation results of the algorithms on the NCUBE/seven hypercube architecture.This work was supported by National Science Foundation grant CCR-8813493. This work was partly done during the author's visit to the Mathematical Science Section, Engineering Physics and Mathematics Division, Oak Ridge National Laboratory, while participating in the Special Year on Numerical Linear Algebra, 1988, sponsored by the UTK Departments of Computer Science and Mathematics, and the ORNL Algebra sponsored by the UTK Departments of Computer Science and Mathematics, and the ORNL Mathematical Sciences Section, Engineering Physics and Mathematics Division.     

实规范矩阵的特征值问题

实对称矩阵的特征值问题,无论是低阶稠密矩阵的全部特征值问题,或高阶稀疏矩阵的部分特征值问题,都已有许多有效的计算方法,迄今最重要的一些成果已总结在[5]中。本文利用规范矩阵的一些重要性质将对于Hermite矩阵(特别是对弥矩阵)特征值问题的一些有效算法推广到规范矩阵的特征值问题,由于对复规范阵的推广是简单的,而且实际上常遇到的是实矩阵(这时常要求只用实运算),因此我们着重讨论实规范矩阵的特征值问题。     

Decomposition of an Updated Correlation Matrix via Hyperbolic Transformations

An algorithm for hyperbolic singular value decomposition of a given complex matrix based on hyperbolic Householder and Givens transformation matrices is described in detail. The main application of this algorithm is the decomposition of an updated correlation matrix.     

矩阵方程AXB=C最小二乘D反对称解

首先将对称矩阵推广到D反对称矩阵,然后研究了方程AXB=C的D反对称最小二乘解,利用矩阵对的广义奇异分解、标准相关分解及子空间上的投影定理,得到了最小二乘解的通式．     

Quadratic convergence of a special quasi-cyclic Jacobi method

This paper considers the ultimate asymptotic convergence of a block- oriented, quasi-cyclic Jacobi method for symmetric matrices. The conclusion applies to the new one-sided Jacobi method for computing the singular value decomposition, recently proposed by Drmač and Veselić. Using a simple qualitative analysis, the discussion indicates that a quadratic off-norm reduction per quasi-sweep is to be expected in all perceivable cases.      

Accurate computations with collocation matrices of a general class of bases

In this paper, we provide algorithms for computing the bidiagonal decomposition of the collocation matrices of a very general class of bases of interest in computer‐aided geometric design and approximation theory. It is also shown that these algorithms can be used to perform accurately some algebraic computations with these matrices, such as the calculation of their inverses, their eigenvalues, or their singular values. Numerical experiments illustrate the results.     

Exploiting Mixed Precision for Computing Eigenvalues of Symmetric Matrices and Singular Values

The use of higher precision preconditioning for the symmetric eigenvalue problem and the singular value problem of general non–structured non–graded matrices are discussed. The matrix Q from the QR–decomposition as a preconditioner, applied to A with higher precision, in combination with Jacobi's method seems to allow the computation of all eigenvalues of symmetric positive definite matrices rsp. all singular values of general matrices to nearly full accuracy. (© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)     

An algorithm for approximating the singular triplets of complex symmetric matrices

We present an algorithm for the approximation of the dominant singular values and corresponding right and left singular vectors of a complex symmetric matrix. The method is based on two short-term recurrences first proposed by Saunders, Simon and Yip [24] for a non-Hermitian linear system solver. With symmetric matrices, the recurrence can be modified so as to generate a tridiagonal symmetric matrix from which the original triplets can be approximated. The recurrence formally resembles the Lanczos method, in spite of substantial differences which make usual convergence results inapplicable. Implementation aspects are discussed, such as re-orthogonalization and the use of alternative representation matrices. The method is very efficient over existing approaches which do not exploit the symmetry of the problem. Numerical experiments on application problems validate the analysis, while showing satisfactory results, especially on dense matrices. © 1997 by John Wiley & Sons, Ltd.     

矩阵方程AXAT+BYBT=C的对称与反对称最小范数最小二乘解

对于任意给定的矩阵A∈Rk×m,B∈Rk×n和C∈Rk×k,利用奇异值分解和广义奇异值分解,我们给出了矩阵方程AXAT+BYBT=C的对称与反对称最小范数最小二乘解的表达式.     

Extensions of Faddeev’s algorithms to polynomial matrices

Starting from algorithms introduced in [Ky M. Vu, An extension of the Faddeev’s algorithms, in: Proceedings of the IEEE Multi-conference on Systems and Control on September 3-5th, 2008, San Antonio, TX] which are applicable to one-variable regular polynomial matrices, we introduce two dual extensions of the Faddeev’s algorithm to one-variable rectangular or singular matrices. Corresponding algorithms for symbolic computing the Drazin and the Moore-Penrose inverse are introduced. These algorithms are alternative with respect to previous representations of the Moore-Penrose and the Drazin inverse of one-variable polynomial matrices based on the Leverrier-Faddeev’s algorithm. Complexity analysis is performed. Algorithms are implemented in the symbolic computational package MATHEMATICA and illustrative test examples are presented.     

A GPU-based hyperbolic SVD algorithm

A one-sided Jacobi hyperbolic singular value decomposition (HSVD) algorithm, using a massively parallel graphics processing unit (GPU), is developed. The algorithm also serves as the final stage of solving a symmetric indefinite eigenvalue problem. Numerical testing demonstrates the gains in speed and accuracy over sequential and MPI-parallelized variants of similar Jacobi-type HSVD algorithms. Finally, possibilities of hybrid CPU–GPU parallelism are discussed.     

Least-Squares Solutions of the Matrix Equation ATXA = B Over Bisymmetric Matrices and its Optimal Approximation

A real n×n symmetric matrix X=(x_(ij))_(n×n)is called a bisymmetric matrix if x_(ij)=x_(n 1-j,n 1-i).Based on the projection theorem,the canonical correlation de- composition and the generalized singular value decomposition,a method useful for finding the least-squares solutions of the matrix equation A~TXA=B over bisymmetric matrices is proposed.The expression of the least-squares solutions is given.Moreover, in the corresponding solution set,the optimal approximate solution to a given matrix is also derived.A numerical algorithm for finding the optimal approximate solution is also described.     

Accurate computations of matrices with bidiagonal decomposition using methods for totally positive matrices

A class of sign‐symmetric P‐matrices including all nonsingular totally positive matrices and their inverses as well as tridiagonal nonsingular H‐matrices is presented and analyzed. These matrices present a bidiagonal decomposition that can be used to obtain algorithms to compute with high relative accuracy their singular values, eigenvalues, inverses, or their LDU factorization. Copyright © 2012 John Wiley & Sons, Ltd.     

The computation of the square roots of circulant matrices

In the first part of this paper, we investigate the reduced forms of circulant matrices and quasi-skew circulant matrices. By using their properties we present two efficient algorithms to compute the square roots of circulant matrices and quasi-skew circulant matrices, respectively. Those methods are faster than the traditional algorithm which is based on the Schur decomposition. In the second part, we further consider circulant H-matrices with positive diagonal entries and develop two algorithms for computing their principal square roots. Those two algorithms have the common advantage that is they only need matrix-matrix multiplications in their iterative sequences, an operation which can be done very efficiently on modern high performance computers.     

Lanczos method for large-scale quaternion singular value decomposition

In many color image processing and recognition applications, one of the most important targets is to compute the optimal low-rank approximations to color images, which can be reconstructed with a small number of dominant singular value decomposition (SVD) triplets of quaternion matrices. All existing methods are designed to compute all SVD triplets of quaternion matrices at first and then to select the necessary dominant ones for reconstruction. This way costs quite a lot of operational flops and CPU times to compute many superfluous SVD triplets. In this paper, we propose a Lanczos-based method of computing partial (several dominant) SVD triplets of the large-scale quaternion matrices. The partial bidiagonalization of large-scale quaternion matrices is derived by using the Lanczos iteration, and the reorthogonalization and thick-restart techniques are also utilized in the implementation. An algorithm is presented to compute the partial quaternion singular value decomposition. Numerical examples, including principal component analysis, color face recognition, video compression and color image completion, illustrate that the performance of the developed Lanczos-based method for low-rank quaternion approximation is better than that of the state-of-the-art methods.