The \(\varepsilon\)-algorithm and \(\eta\)-algorithm for generalized inverse rectangular matrix Padé approximants (Q4515595)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 1529394
Language Label Description Also known as
default for all languages
No label defined
    English
    The \(\varepsilon\)-algorithm and \(\eta\)-algorithm for generalized inverse rectangular matrix Padé approximants
    scientific article; zbMATH DE number 1529394

      Statements

      The ε-algorithm and η-algorithm for generalized inverse rectangular matrix pade approximants (English)
      0 references
      The \(\varepsilon\)-algorithm and \(\eta\)-algorithm for generalized inverse rectangular matrix Padé approximants (English)
      0 references
      0 references
      23 November 2001
      0 references
      matrix valued Padé approximants
      0 references
      \(\varepsilon\)-algorithm
      0 references
      \(\eta\)-algorithm
      0 references
      \textit{C. Gu}, \textit{Z. Chen} [Math. Numer. Sin. 17, No. 1, 73-77 (1995; Zbl 0860.41018)] gave an axiomatic introduction of matrix Padé approximants using a generalized inverse in the vein of the Moore-Penrose inverse for a vector. This makes it possible to define the Padé approximants without resorting to making a difference between left and right approximant. The inverse of \(A=(a_{ij})^n_{i,j=1}\) is given as follows: \(A^{-1}=A^{*}/||A||^2,\) where \(||A||^2=\sum_i\sum_j |a_{ij}|^2\) and \({}^{*}\) denotes complex conjugation. In this paper the author defines a matrix valued \(\varepsilon\)-algorithm (a copy of the ordinary one due to P. Wynn), leading to a two-dimensional table that can be used to construct the Matrix Padé table. This algorithm also satisfies the famous Wynn-identity \[ (N-C)^{-1}+(S-C)^{-1}=(E-C)^{-1}+(W-C)^{-1}, \] where N, S, E, W and C mean North, South, East, West and Center. Furthermore a matrix \(\eta\)-algorithm and its two-dimensional table are introduced and a numerical example is given. There are several easily corrected typos in the paper, there is a serious problem with formula (2.9) [why is it allowed to leave out the absolute value bars around the \(z^2\)?] and the author does not state explicitly that, when referring to examples in the book by \textit{G. A. Baker jun.} and \textit{P. Graves-Morris} [Padé approximants. I, II (1981; Zbl 0468.30032 and Zbl 0468.30033)], some formulae use the ordinary inverse and not the generalized one (in his example 2.2 is referred to \({}^{\mathbb{R}}[1/1]\): the upper right hand element in the second matrix of the first line should be \(+z/2\) to make it correct).
      0 references

      Identifiers