No label defined (Q4387423)

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 1150010
Language Label Description Also known as
default for all languages
No label defined
    English
    No label defined
    scientific article; zbMATH DE number 1150010

      Statements

      17 November 1998
      0 references
      zeros of polynomials
      0 references
      orthogonal expansions
      0 references
      orthogonal polynomials
      0 references
      inequalities
      0 references
      norm estimates
      0 references
      0 references
      Given a complex \(n\)-th degree polynomial \(p(z)= \sum^n_{\nu=0} a_\nu z^\nu \), it was observed by Cauchy in 1829 that the associated polynomial \(\sum^{n-1}_{\nu=0} | a_\nu| z^\nu -| a_n| z^n\) has exactly one zero \(\rho_n [p]\geq 0\) and that \(| z_0 |\leq \rho_n [p]\) holds for every zero \(z_0\) of \(p(z)\). The author considers polynomials \(f(z)= \sum^n_{\nu=0} a_\nu \Phi_\nu(z)\), where \(\{\Phi_\nu (z)\}^\infty_{\nu=0}\) are real monic polynomials, orthogonal with respect to a positive Borel measure \(\mu\). Denoting by \(J_n\) the smallest compact interval that contains all zeros of \(\Phi_n\), put \(d_n(z): =\min_{t\in J_n} | z-t |\) for \(z\in \mathbb{C}\). Then the author presents a survey of bounds for \(d_n (\zeta)\), where \(\zeta\) is a zero of \(f(z)\). Many of these results were already obtained in 1964 by W. Specht for general orthogonal polynomials and by P. Turán for the special case of Hermite polynomials. One typical results with a surprisingly simple and elegant proof is Theorem 1. If \(f(\zeta)=0\), then \(|\text{Im} (\zeta) |\leq d_n (\zeta)\leq \rho_n[p]\) holds. Since every upper bound for the moduli of all zeros of \(p(z)\), that depends on \(| a_\nu/a_n|\), \(0\leq \nu<n\), only, also is an upper bound for \(\rho_n[p]\), a long list of upper bounds for \(d_n(\zeta)\) is available. Theorem 1 remains true if \(\Phi_\nu\), \(\nu\in \mathbb{N}\), are Newton interpolation polynomials. Another result of the author is theorem 2. If \(x_1, \dots, x_n\) are the zeros of \(\Phi_n\), then every polynomial \(f(z)\) has all its zeros in the union of the disks \(D_\nu: =\{z\in\mathbb{C}: | z-x_\nu |\leq r_n\}\), \(1\leq \nu\leq n\), where \[ r_n: =\left(\sum^{n-1}_{\nu=0} (\gamma_\nu/ \gamma_{\nu-1}) | a_\nu/a_n |^2 \right)^{1/2} \quad \text{with} \quad \gamma_\nu: =\int_{-\infty}^{+\infty} \Phi^2_\nu(x) d\mu(x). \] This yields a result of W. Specht that \(| \text{Im} \zeta |\leq d_n (\zeta)\leq r_n\) holds for every zero \(\zeta\) of \(f(z)\). The author presents many further and more involved estimates including sharper results for real polynomials \(f(z)\) and Hermite polynomials \(\Phi_\nu(z)\). For the entire collection see [Zbl 0882.00013].
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references