Exclusion and inclusion regions for the eigenvalues of a normal matrix (Q2784398)

From MaRDI portal





scientific article; zbMATH DE number 1732288
Language Label Description Also known as
English
Exclusion and inclusion regions for the eigenvalues of a normal matrix
scientific article; zbMATH DE number 1732288

    Statements

    0 references
    0 references
    23 April 2002
    0 references
    Bézout's theorem
    0 references
    algebraic subvariety of \(\mathbb R^2\)
    0 references
    eigenvalue localization
    0 references
    normal matrix
    0 references
    polyanalytic polynomial
    0 references
    slowly growing length of the recurrence
    0 references
    Ritz value
    0 references
    exclusion and inclusion regions
    0 references
    Arnoldi type iteration
    0 references
    Exclusion and inclusion regions for the eigenvalues of a normal matrix (English)
    0 references
    Let \(A\) be an \(n\times n\) normal matrix with characteristic polynomial \(p(z)= f(x,y)+ ig(x,y)\). It is shown that because of the normality of \(A\), real analytic techniques yield bivariate polynomials such that generically, NEWLINE\[NEWLINE\deg f\deg g\leq 4n\tag{1}NEWLINE\]NEWLINE holds pairwise, so that in the light of Bézout's theorem, (1) is of correct order. The corresponding algorithm involves a recurrence with very slowly growing length.NEWLINENEWLINENEWLINEAlternatives for large problems proposed are Ritz calculations and methods for eigenvalue exclusion and inclusion regions. In practice, polyanalytic polynomials needed in the present context are generated by an Arnoldi type iteration.
    0 references

    Identifiers