A class of direct methods for linear systems

Summary A class of methods of direct type for solving determined or underdetermined, full rank or deficient rank linear systems is presented and theoretically analyzed. The class can be considered as a generalization of the methods of Brent and Brown as restricted to linear systems and implicitly contains orthogonal,LU andLL ^T factorization methods.     

A finite element — Capacitance method for elliptic problems on regions partitioned into subregions

Summary A method is given for the solution of linear equations arising in the finite element method applied to a general elliptic problem. This method reduces the original problem to several subproblems (of the same form) considered on subregions, and an auxiliary problem. Very efficient iterative methods with the preconditioning operator and using FFT are developed for the auxiliary problem.     

Algebraic connections between the least squares and total least squares problems

Summary In this paper the closeness of the total least squares (TLS) and the classical least squares (LS) problem is studied algebraically. Interesting algebraic connections between their solutions, their residuals, their corrections applied to data fitting and their approximate subspaces are proven.All these relationships point out the parameters which mainly determine the equivalences and differences between the two techniques. These parameters also lead to a better understanding of the differences in sensitivity between both approaches with respect to perturbations of the data.In particular, it is shown how the differences between both approaches increase when the equationsAXB become less compatible, when the length ofB orX is growing or whenA tends to be rank-deficient. They are maximal whenB is parallel with the singular vector ofA associated with its smallest singular value. Furthermore, it is shown how TLS leads to a weighted LS problem, and assumptions about the underlying perturbation model of both techniques are deduced. It is shown that many perturbation models correspond with the same TLS solution.Senior Research Assistant of the Belgian N.F.W.O. (National Fund of Scientific Research)     

Least squares with a quadratic constraint

Summary We present the theory of the linear least squares problem with a quadratic constraint. New theorems characterizing properties of the solutions are given. A numerical application is discussed.     

Numerical solution of eigentuple-eigenvector problems in Hilbert spaces by a gradient method

Summary A gradient technique previously developed for computing the eigenvalues and eigenvectors of the general eigenproblemAx=Bx is generalized to the eigentuple-eigenvector problem . Among the applications of the latter are (1) the determination of complex (,x) forAx=Bx using only real arithmetic, (2) a 2-parameter Sturm-Liouville equation and (3) -matrices. The use of complex arithmetic in the gradient method is also discussed. Computational results are presented.This research was partially supported by NSF Grants MPS74-13332 and MCS76-09172     

AB-Algorithm and its modifications for the spectral problems of linear pencils of matrices

Summary This paper describes and algorithm and its modifications for solving spectral problems for linear pencils of matrices both regular as well as singular.     

A parallel projection method for solving generalized linear least-squares problems

Summary A parallel projection algorithm is proposed to solve the generalized linear least-squares problem: find a vector to minimize the 2-norm distance from its image under an affine mapping to a closed convex cone. In each iteration of the algorithm the problem is decomposed into several independent small problems of finding projections onto subspaces, which are simple and can be tackled parallelly. The algorithm can be viewed as a dual version of the algorithm proposed by Han and Lou [8]. For the special problem under consideration, stronger convergence results are established. The algorithm is also related to the block iterative methods of Elfving [6], Dennis and Steihaug [5], and the primal-dual method of Springarn [14].This material is based on work supported in part by the National Science foundation under Grant DMS-8602416 and by the Center for Supercomputing Research and Development, University of Illinois at Urbana-Champaign     

A Newton iteration process for inverse eigenvalue problems

Summary We suppose an inverse eigenvalue problem which includes the classical additive and multiplicative inverse eigenvalue problems as special cases. For the numerical solution of this problem we propose a Newton iteration process and compare it with a known method. Finally we apply it to a numerical example.     

A decomposition-dualization approach for solving constrained convex minimization problems with applications to discretized obstacle problems

Summary In this paper, we shall be concerned with the solution of constrained convex minimization problems. The constrained convex minimization problems are proposed to be transformable into a convex-additively decomposed and almost separable form, e.g. by decomposition of the objective functional and the restrictions. Unconstrained dual problems are generated by using Fenchel-Rockafellar duality. This decomposition-dualization concept has the advantage that the conjugate functionals occuring in the derived dual problem are easily computable. Moreover, the minimum point of the primal constrained convex minimization problem can be obtained from any maximum point of the corresponding dual unconstrained concave problem via explicit return-formulas. In quadratic programming the decomposition-dualization approach considered here becomes applicable if the quadratic part of the objective functional is generated byH-matrices. Numerical tests for solving obstacle problems in ¹ discretized by using piecewise quadratic finite elements and in ² by using the five-point difference approximation are presented.     

A new method for the solution ofAx=b

Summary An alternative to Gauss elimination is developed to solveAx=b. The method enables one to exploit zeros in the lower triangle ofA, so that the number of multiplications needed to perform the algorithm can be substantially reduced.     

Alternating methods for sets of linear equations

Summary A generalization of alternating methods for sets of linear equations is described and the number of operations calculated. It is shown that the lowest number of arithmetic operations is achieved in the SSOR algorithm.     

On the convergence of the conjugate gradient method for singular capacitance matrix equations from the Neumann problem of the Poisson equation

Summary It is shown analytically in this work that the conjugate gradient method is an efficient means of solving the singular capacitance matrix equations arising from the Neumann problem of the Poisson equation. The total operation count of the algorithm does not exceed constant timesN ²logN(N=1/h) for any bounded domain with sufficiently smooth boundary.Sponsored by the United States Army under Contract No. DAAG29-75-C-0024 and the National Science Foundation under Grant No. MCS75-17385. Also partially supported by the Energy Research and Development Administration     

Round-off error analysis of iterations for large linear systems

Summary We deal with the rounding error analysis of successive approximation iterations for the solution of large linear systemsA _x =b. We prove that Jacobi, Richardson, Gauss-Seidel and SOR iterations arenumerically stable wheneverA=A ^*>0 andA has PropertyA. This means that the computed resultx _k approximates the exact solution with relative error of order A·A ^–1 where is the relative computer precision. However with the exception of Gauss-Seidel iteration the residual vector Ax _k –b is of order A² A ^–1 and hence the remaining three iterations arenot well-behaved.This work was partly done during the author's visit at Carnegie-Mellon University and it was supported in part by the Office of Naval Research under Contract N00014-76-C-0370; NR 044-422 and by the National Science Foundation under Grant MCS75-222-55     

A divide and conquer method for the symmetric tridiagonal eigenproblem

Summary A method is given for calculating the eigenvalues of a symmetric tridiagonal matrix. The method is shown to be stable and for a large class of matrices it is, asymptotically, faster by an order of magnitude than theQR method.     

Convergent nonnegative matrices and iterative methods for consistent linear systems

Summary In this paper we study linear stationary iterative methods with nonnegative iteration matrices for solving singular and consistent systems of linear equationsAx=b. The iteration matrices for the schemes are obtained via regular and weak regular splittings of the coefficients matrixA. In certain cases when only some necessary, but not sufficient, conditions for the convergence of the iterations schemes exist, we consider a transformation on the iteration matrices and obtain new iterative schemes which ensure convergence to a solution toAx=b. This transformation is parameter-dependent, and in the case where all the eigenvalues of the iteration matrix are real, we show how to choose this parameter so that the asymptotic convergence rate of the new schemes is optimal. Finally, some applications to the problem of computing the stationary distribution vector for a finite homogeneous ergodic Markov chain are discussed.Research sponsored in part by US Army Research Office     

On the global convergence of the Eberlein method for real matrices

Summary The global convergence proof of the column-and row-cyclic Eberlein diagonalization process for real matrices is given. The convergence to a fixed matrix in Murnaghan form is obtained with the well-known exception of complex-conjugate pairs of eigenvalues whose real parts are more than double.     

On the convergence of the symmetric SOR method for matrices with red-black ordering

Summary We prove that if the matrixA has the structure which results from the so-called red-black ordering and ifA is anH-matrix then the symmetric SOR method (called the SSOR method) is convergent for 0<<2. In the special case thatA is even anM-matrix we show that the symmetric single-step method cannot be accelerated by the SSOR method. Symmetry of the matrixA is not assumed.     

Precise domains of convergence for the block SSOR method associated withp-cyclic matrices

We apply Rouché's theorem to the functional equation relating the eigenvalues of theblock symmetric successive overrelaxation (SSOR) matrix with those of the block Jacobi iteration matrix found by Varga, Niethammer, and Cai, in order to obtain precise domains of convergence for the block SSOR iteration method associated withp-cyclic matricesA, p3. The intersection of these domains, taken over all suchp's, is shown to coincide with the exact domain of convergence of thepoint SSOR iteration method associated withH-matricesA. The latter domain was essentially discovered by Neumaier and Varga, but was recently sharpened by Hadjidimos and Neumann.Research supported in part by NSF Grant DMS 870064.     

Fehleranalyse für die Gauß-Elimination zur Berechnung der Lösung minimaler Länge

Summary Presented is a realistic, elementwise analysis for the rounding errors of a generalization of Gauss elimination for solving the linear best least squares problem without pivoting. The bounds are suitable to determine the class of well-posed problems to the given method. A mixed error analysis is given and then the effects of errors in the input data are studied. Numerical examples demonstrate the efficiency.

     

Orthogonal least squares fitting with linear manifolds

Summary The problem is considered of fitting a linear manifold of dimensions with 1sn–1 to a given set of points in ⁿ such that the sum of orthogonal squared distances attains a minimum.