The generalized Levinger transformation (Q848560)

From MaRDI portal





scientific article
Language Label Description Also known as
default for all languages
No label defined
    English
    The generalized Levinger transformation
    scientific article

      Statements

      The generalized Levinger transformation (English)
      0 references
      4 March 2010
      0 references
      New results relating the numerical range of a matrix \(A\) with the generalized Levinger transformation \[ \mathcal L(A,\alpha ,\beta) = \alpha H_A +\beta S_A = \frac{\alpha+\beta}2A+ \frac{\alpha-\beta}2A^* \] are presented; here \(H_A\) and \(S_A\) are the Hermitian and skew-Hermitian parts of \(A\), respectively. This relationship is used to obtain results on eigenvalues and eigenvectors of the perturbed matrix of the form \(A+\mathcal L(E,\alpha ,\beta)\), where \(E\) is a fixed matrix, and \(\alpha, \beta \) are small real parameters. The first part of the paper contains geometric properties of numerical range of \(\mathcal L(A,\alpha,\beta) \) and bounds are given for the real and imaginary parts of eigenvalues of \(\mathcal L(A,\alpha,\beta) \). In the second part, the Levinger transformation is applied to a given fixed matrix \(E\), viewed as a perturbation matrix, whose activity on a matrix \(A\) depends only on the parameters \(\alpha\) and \(\beta\). A necessary and sufficient condition for a normal matrix to remain normal under a perturbation by a symmetric and rank one matrix is given, and an~approximation of a perturbed eigenpair of a diagonalizable matrix \(A\) in terms of two parameters is considered. As an application, a sufficient condition for the perturbed eigenpair to be first order approximation of the corresponding simple eigenpair of initial matrix \(A\) is given, and two numerical examples to illustrate the results obtained are presented.
      0 references
      numerical range
      0 references
      generalized inverses
      0 references
      perturbation theory
      0 references
      eigenvalues
      0 references
      eigenvectors
      0 references
      field of values
      0 references
      control and systems theory
      0 references
      sensitivity
      0 references
      Levinger transformation
      0 references
      normal matrix
      0 references
      numerical examples
      0 references
      0 references
      0 references
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references