Weighted polynomial approximation for convex external fields (Q1570053)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Weighted polynomial approximation for convex external fields |
scientific article |
Statements
Weighted polynomial approximation for convex external fields (English)
0 references
9 July 2000
0 references
In this landmark paper, Vilmos Totik largely settles one of the problems that has intrigued researchers in orthogonal polynomials, extremal polynomials, exponential weights, \(\dots\). Let \(\Sigma\subset\mathbb{R}\) and \(Q: \Sigma \to\mathbb{R}\). What functions may be uniformly approximated on \(\Sigma\) by weighted polynomials of the form \(P_n\exp (-nQ)\) as \(n\to\infty\)? In this question, potential theory for external fields plays a crucial role. It is known that there is a certain equilibrium measure \(\mu_Q\) associated with \(Q\), and with support \(S_Q\) in \(\Sigma\), that governs the solution of this problem. If \(f:\sum \to\mathbb{R}\) and there exist polynomials \(P_n\) of degree \(\geq n\) such that \[ \lim_{n\to \infty}\bigl \|f-P_n\exp(-nQ)\bigr \|_{L_\infty (\Sigma)}=0, \] then \(f\) must vanish outside \(S_Q\). The author essentially settles the converse for convex \(Q\): if \((a,b)\) is contained in \(S_Q\) and \(Q\) is convex on \((a,b)\), then every function \(f\) vanishing outside \((a,b)\) may be uniformly approximated on \(\Sigma\) by weighted polynomials of the form \(P_n\exp (-nQ)\). This follows from more general results about equilibrium measures with ``smooth integral''. The methods of proof confirm that Totik is a technical virtuoso; he uses his method of very careful discretisation of a potential in a manner that only he could have invented. The paper is essential reading for those involved in potential theory for external fields, weighted approximation, orthogonal polynomials for exponential weights,\(\dots\).
0 references
exponential weights
0 references
potential theory
0 references
weighted approximation
0 references