Backward Error and Condition of Structured Linear Systems

Existing definitions of backward error and condition number for linear systems do not cater to structure in the coefficient matrix, except possibly for sparsity. The definitions are extended so that when the coefficient matrix has structure the perturbed matrix has this structure too. It is shown that when the structure comprises linear dependence on a set of parameters, the structured componentwise backward error is given by the solution of minimal $\infty $ -norm to an underdetermined linear system; an explicit expression for the condition number in this linear case is also obtained. Applications to symmetric matrices, Toeplitz matrices and the least squares problem are discussed and illustrated through numerical examples.

[1]  Gene H. Golub,et al.  Numerical Linear Algebra, Digital Signal Processing and Parallel Algorithms , 1991, NATO ASI Series.

[2]  W. Prager,et al.  Compatibility of approximate solution of linear equations with given error bounds for coefficients and right-hand sides , 1964 .

[3]  J. Bunch.  The weak and strong stability of algorithms in numerical linear algebra , 1987 .

[4]  James Hardy Wilkinson,et al.  Rounding errors in algebraic processes , 1964, IFIP Congress.

[5]  N. Higham.  Computing error bounds for regression problems , 1990 .

[6]  D SkeelRobert.  Scaling for Numerical Stability in Gaussian Elimination , 1979 .

[7]  J. Demmel,et al.  The strong stability of algorithms for solving symmetric linear systems , 1989 .

[8]  Robert D. Skeel,et al.  Scaling for Numerical Stability in Gaussian Elimination , 1979, JACM.

[9]  G. Watson.  Approximation theory and numerical methods , 1980 .

[10]  F. R. Gantmakher.  The Theory of Matrices , 1984 .

[11]  Paul Van Dooren,et al.  Structured linear algebra problems in digital signal processing , 1991 .

[12]  N. Higham.  Iterative refinement enhances the stability ofQR factorization methods for solving linear equations , 1991 .

[13]  J. H. Wilkinson.  The algebraic eigenvalue problem , 1966 .

[14]  J. L. Rigal,et al.  On the Compatibility of a Given Solution With the Data of a Linear System , 1967, JACM.

[15]  Nicholas J. Higham,et al.  FORTRAN codes for estimating the one-norm of a real or complex matrix, with applications to condition estimation , 1988, TOMS.

[16]  Nicholas J. Higham,et al.  Experience with a Matrix Norm Estimator , 1990, SIAM J. Sci. Comput..

[17]  Nicholas J. Higham,et al.  Error analysis of the Björck-Pereyra algorithms for solving Vandermonde systems , 1987 .

[18]  John B. Shoven,et al.  I , Edinburgh Medical and Surgical Journal.

[19]  Å. Björck.  Component-wise perturbation analysis and error bounds for linear least squares solutions , 1991 .

[20]  T. Coleman,et al.  A global and quadratically convergent method for linear l ∞ problems , 1992 .

[21]  Charles R. Johnson.  The Theory of Matrices. Second Edition (with Applications) (Peter Lancaster and Miron Tismenetsky) , 1987 .

[22]  I. Duff,et al.  On the augmented system approach to sparse least-squares problems , 1989 .

[23]  Peter Lancaster,et al.  The theory of matrices , 1969 .