On a conjecture of Widom

DOI10.1016/J.JAT.2014.04.006MaRDI QIDQ2514841zbMATH OpenOpenAlexWikidataFDO

Publication date 4 February 2015

Published in Journal of Approximation Theory (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1401.6357

Chebyshev polynomials Jordan arcs supremum norm Widom's theory

Potentials and capacity, harmonic measure, extremal length and related notions in two dimensions (31A15) Orthogonal functions and polynomials, general theory of nontrigonometric harmonic analysis (42C05)

Abstract: In 1969 Harold Widom published his seminal paper, which gave a complete description of orthogonal and Chebyshev polynomials on a system of smooth Jordan curves. When there were Jordan arcs present the theory of orthogonal polynomials turned out to be just the same, but for Chebyshev polynomials Widom's approach proved only an upper estimate, which he conjectured to be the correct asymptotic behavior. In this note we make some clarifications which will show that the situation is more complicated.

Review text

Let \(E\) be the union of finitely many disjoint Jordan curves or arcs \(E_{k}\) that lie in the exterior of each other and that satisfy some smoothness condition. Given a nonnegative weight function on \(E\), the (weighted) Chebyshev polynomial of degree \(n\) associated with \(E\) is defined as the unique monic polynomial \(T_{n, \rho}(z)\) which minimizes the supremum norm \[ \sup_{z \in E} |Q_{n}(z)| \rho(z), \] with minimal norms notated by \(M_{n, \rho}\). \newline Being already known that, when \(\rho \equiv 1\), the logarithmic capacity \(C(E)\) allows the definition of a lower bound for \(M_{n}:=M_{n,1}\), through the inequality \[ M_{n} \geq C(E)^{n}, \; n=1,2,\ldots, \] this contribution revisits \textit{H. Widom}'s paper [Adv. Math. 3, 127--232 (1969; Zbl 0183.07503)], and deals with asymptotic formulas regarding the Chebyshev numbers \(M_{n,\rho}\). Assuming \(E\) only formed by Jordan curves, Widom established the following asymptotic behavior for \(M_{n,\rho}\): \[ M_{n,\rho} \sim C(E)^{n} \mu\left( \rho, \Gamma_{n} \right), \] that is, the ratio between \(M_{n,\rho}\) and \(C(E)^{n} \mu\left( \rho, \Gamma_{n} \right)\) tends to \(1\), for certain quantities \(\mu \left( \rho, \Gamma_{n} \right)\) therein marked out. \newline A concise description of Widom's paper is given, recalling that when \(E\) has Jordan arcs, Widom left incomplete the characterization of Chebyshev polynomials and numbers, although he conjectured that, having \(E\) a Jordan arc component, the formula proved for \(M_{n,\rho}\) in the Jordan curves situation should be multiplied by 2. The articulation between the content of Widom's paper and the ideas here developed allows a better understanding of the difficulties of the problem and of the progress achieved. In particular, the authors found in Widom's work the equipment needed to clarify several aspects regarding that conjecture, which is shown to be false (in its strict sense). Indeed, they prove the following result concerned with the case where \(E\) has both curve and arc components: ``If there is at least one Jordan curve in \(E\), then \[ \limsup_{n \to \infty} \frac{M_{n,\rho}}{C(E)^{n}\mu\left( \rho, \Gamma_{n} \right)} \leq \theta <2 \] where \(\theta\) depends only on \(E\). Supposing that \(E\) is composed by real intervals and Jordan curves symmetric with respect to the real line, it is shown that \(M_{n,\rho} \sim C(E)^{n} \mu\left( \rho^{\ast}, \Gamma_{n} \right)\), where \(\rho^{\ast}\) is equal to \(\rho\) on the curve components of \(E\) and equal to \(2\rho\) on the arc components. The authors also provide an explicit formula for the asymptotics of \(M_{n,1}\) assuming that \(E\) consists of an interval \(\left[\alpha_{1}, \beta_{1}\right]\) and a symmetric Jordan curve \(E_{2}\).

Recommendations

Cites work

Cited in

(18)

This page was built for publication: On a conjecture of Widom

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2514841)