Sensitivity analysis for the generalized singular value decomposition

In this paper, we discuss the sensitivity of multiple nonzero finite generalized singular values and the corresponding generalized singular matrix set of a real matrix pair analytically dependent on several parameters. From our results, the partial derivatives of multiple nonzero singular values and their left and right singular vector matrices are obtained.Copyright © 2012 John Wiley & Sons, Ltd.     

The generalized triangular decomposition

Given a complex matrix , we consider the decomposition , where is upper triangular and and have orthonormal columns. Special instances of this decomposition include the singular value decomposition (SVD) and the Schur decomposition where is an upper triangular matrix with the eigenvalues of on the diagonal. We show that any diagonal for can be achieved that satisfies Weyl's multiplicative majorization conditions:

where is the rank of , is the -th largest singular value of , and is the -th largest (in magnitude) diagonal element of . Given a vector which satisfies Weyl's conditions, we call the decomposition , where is upper triangular with prescribed diagonal , the generalized triangular decomposition (GTD). A direct (nonrecursive) algorithm is developed for computing the GTD. This algorithm starts with the SVD and applies a series of permutations and Givens rotations to obtain the GTD. The numerical stability of the GTD update step is established. The GTD can be used to optimize the power utilization of a communication channel, while taking into account quality of service requirements for subchannels. Another application of the GTD is to inverse eigenvalue problems where the goal is to construct matrices with prescribed eigenvalues and singular values.

     

广义逆A_T~(2),S的一个特征

给出了矩阵广义逆AT,B(2)的一个特征,并由此建立了AT,B(2)的一种算法.     

矩阵的加权Moore—Penrose逆

利用矩阵的广义奇异值分解，给出了复数域上矩阵的Moore—Penrose逆存在的充要条件及其表达式．     

四元数矩阵的奇异值分解及其应用

In this paper, a constructive proof of singular value decomposition of quaternion matrix is given by using the complex representation and companion vector of quaternion matrix and the computational method is described. As an application of the singular value decomposition, the CS decomposition is proved and the canonical angles on subspaces of Q^n is studied.     

矩阵方程AXB=D的最小二乘Hermite解及其加权最佳逼近

本中,我们讨论了矩阵方程AXB=D的最小二乘Hermite解,通过运用广义奇异值分解(GSVD)，获得了解的通式。此外,对于给定矩阵F，也得到了它的加权最佳逼近表达式。     

A QR DECOMPOSITION BASED SOLVER FOR THE LEAST SQUARES PROBLEMS FROM THE MINIMAL RESIDUAL METHOD FOR THE SYLVESTER EQUATION

Based on the generalized minimal residual （GMRES） principle, Hu and Reichel proposed a minimal residual algorithm for the Sylvester equation. The algorithm requires the solution of a structured least squares problem. They form the normal equations of the least squares problem and then solve it by a direct solver, so it is susceptible to instability. In this paper, by exploiting the special structure of the least squares problem and working on the problem directly, a numerically stable QR decomposition based algorithm is presented for the problem. The new algorithm is more stable than the normal equations algorithm of Hu and Reichel. Numerical experiments are reported to confirm the superior stability of the new algorithm.     

关于四元数矩阵的最佳逼近问题

通过使用四元数矩阵的广义奇异值分解,给出了四元数矩阵最佳逼近问题‖AHXA-C‖2F+‖BHXB-D‖2F=min, s.t. XH=X的一般表达式.     

MINIMIZATION PROBLEM FOR SYMMETRIC ORTHOGONAL ANTI-SYMMETRIC MATRICES

By applying the generalized singular value decomposition and the canonical correlation decomposition simultaneously, we derive an analytical expression of the optimal approximate solution ^-X, which is both a least-squares symmetric orthogonal anti-symmetric solu- tion of the matrix equation A^TXA = B and a best approximation to a given matrix X^＊. Moreover, a numerical algorithm for finding this optimal approximate solution is described in detail, and a numerical example is presented to show the validity of our algorithm.     

Finite difference scheme based on proper orthogonal decomposition for the nonstationary Navier-Stokes equations

The proper orthogonal decomposition(POD)and the singular value decomposition(SVD) are used to study the finite difference scheme(FDS)for the nonstationary Navier-Stokes equations. Ensembles of data are compiled from the transient solutions computed from the discrete equation system derived from the FDS for the nonstationary Navier-Stokes equations.The optimal orthogonal bases are reconstructed by the elements of the ensemble with POD and SVD.Combining the above procedures with a Galerkin projection approach yields a new optimizing FDS model with lower dimensions and a high accuracy for the nonstationary Navier-Stokes equations.The errors between POD approximate solutions and FDS solutions are analyzed.It is shown by considering the results obtained for numerical simulations of cavity flows that the error between POD approximate solution and FDS solution is consistent with theoretical results.Moreover,it is also shown that this validates the feasibility and efficiency of POD method.     

酉延拓矩阵的奇异值分解及其广义逆

从普通奇异值分解出发,导出了酉延拓矩阵的奇异值和奇异向量与母矩阵的奇异值和奇异向量间的定量关系,同时对酉延拓矩阵的满秩分解及g逆,反射g逆,最小二乘g逆,最小范数g逆作了定量分析,得到了酉延拓矩阵的满秩分解矩阵F*和G*与母矩阵A的分解矩阵F和G之间的关系.最后给出了相应的快速求解算法,并举例说明该算法大大降低了分解的计算量和存储量,提高了计算效率.     

Singular value decomposition expansion for electrical demand analysis

Email: fan.li{at}ngc.co.uk To analyse electricity-demand data, appropriate mathematicalmodels and algorithms have been developed. Some theoreticalproperties of the singular value decomposition (SVD) and SVDexpansion have been found very useful during these developments.These properties, which have generic implications in data miningand numerical analysis, are presented in this paper. Followinga discussion of the theoretical development, this paper reportsapplications of the SVD expansion in demand analysis and forecastingwith three illustrative practical algorithms, that have beendeveloped by the National Grid Company in recent years.     

‘Fat’ triangulations,or solving certain nonconvex matrix optimization problems

In an earlier paper, the author proposed the problems of determining ‘optimal’ linear transformations of the triangulationsJ ₁ andK ₁, in the sense of minimizing their average directional density for a given mesh size. These tasks were also formulated as optimization problems where the variable is a matrix. Here we solve these problems, and another one which is analogously related to finding an ‘optimal’ linear transformation of the new triangulationJ′. We show thatJ ₁ andJ′ are themselves optimal, while the (α*β*) ofK ₁ developed by van der Laan and Talman is optimal. The latter theorem extends partial results of van der Laan and Talman and Eaves. The optimality of these linear transformations is quite robust: we may change the objective function to maximizing the volume of each simplex, or the constraints to limiting the sum of squares of edge lengths of each simplex, or both, without changing the optimal solutions. Research partially supported by a fellowship from the Alfred P. Sloan Foundation and by NSF Grant ENG82-15361     

A modified truncated singular value decomposition method for discrete ill‐posed problems

Truncated singular value decomposition is a popular method for solving linear discrete ill‐posed problems with a small to moderately sized matrix A. Regularization is achieved by replacing the matrix A by its best rank‐k approximant, which we denote by A_k. The rank may be determined in a variety of ways, for example, by the discrepancy principle or the L‐curve criterion. This paper describes a novel regularization approach, in which A is replaced by the closest matrix in a unitarily invariant matrix norm with the same spectral condition number as A_k. Computed examples illustrate that this regularization approach often yields approximate solutions of higher quality than the replacement of A by A_k.Copyright © 2014 John Wiley & Sons, Ltd.     

Hamiltonian decomposition of complete bipartiter-hypergraphs

In [1] the concepts of paths and cycles of a hypergraph were introduced. In this paper, we give the concepts for bipartite hypergraph and Hamiltonian paths and cycles of a hypergraph, and prove that the complete bipartite 3-hypergraph withq vertices in each part is Hamiltonian decomposable whereq is a prime. This research is supported by the National Natural Science Foundation of China (No.19831080).     

A simultaneous decomposition of a matrix triplet with applications

Researches on ranks of matrix expressions have posed a number of challenging questions, one of which is concerned with simultaneous decompositions of several given matrices. In this paper, we construct a simultaneous decomposition to a matrix triplet (A, B, C), where A=±A*. Through the simultaneous matrix decomposition, we derive a canonical form for the matrix expressions A?BXB*?CYC* and then solve two conjectures on the maximal and minimal possible ranks of A?BXB*?CYC* with respect to X=±X* and Y=±Y*. As an application, we derive a sufficient and necessary condition for the matrix equation BXB* + CYC*=A to have a pair of Hermitian solutions, and then give the general Hermitian solutions to the matrix equation. Copyright © 2010 John Wiley & Sons, Ltd.     

OPTIMAL APPROXIMATE SOLUTION OF THE MATRIX EQUATION AXB = C OVER SYMMETRIC MATRICES

Let SE denote the least-squares symmetric solution set of the matrix equation A×B = C, where A, B and C are given matrices of suitable size. To find the optimal approximate solution in the set SE to a given matrix, we give a new feasible method based on the projection theorem, the generalized SVD and the canonical correction decomposition.     

A randomized tensor singular value decomposition based on the t‐product

The tensor SVD (t‐SVD) for third‐order tensors, previously proposed in the literature, has been applied successfully in many fields, such as computed tomography, facial recognition, and video completion. In this paper, we propose a method that extends a well‐known randomized matrix method to the t‐SVD. This method can produce a factorization with similar properties to the t‐SVD, but it is more computationally efficient on very large data sets. We present details of the algorithms and theoretical results and provide numerical results that show the promise of our approach for compressing and analyzing image‐based data sets. We also present an improved analysis of the randomized and simultaneous iteration for matrices, which may be of independent interest to the scientific community. We also use these new results to address the convergence properties of the new and randomized tensor method as well.