Convergence of polynomials with restricted zeros

From MaRDI portal





This is an elementary exposition of two topics in approximation theory. The author considers the (very ) classical situation that a function \(f(x), x \in I = [0,1]\) can be uniformly approximated by polynomials with real zeros (resp. zeros in the closed upper half-plane). The conclusion in each case is that \(f\) is the restriction to I of an entire function of Laguerre-Pólya class (resp. Pólya-Obrechkoff class). These are well- known results, which are referenced in the paper, but the exposition herein is very elementary and elegant, making heavy use of normal family techniques. Here is one of the two ideas used in the L-P case. We may assume that \(f > m > 0\) on I. Then if \(\{p_ n\}\) are the approximating polynomials, the condition that the zeros be real ensures that \(\{p_ n^{-1}\}\) converge in the strip \(\{\Re z < 1 \}\). Then by an elementary analysis, the author concludes that the \(\{p_ n\}\) converge normally in the full plane, from which it is immediate that \(f \in L-P\). The P-O class is reduced to the L-P case by considering real parts of the approximating polynomials. For the entire collection see [Zbl 0745.00065].











This page was built for publication: Convergence of polynomials with restricted zeros

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4027220)