Table of Contents
Fetching ...

Nehari's Theorem and Hardy's inequality for Paley--Wiener spaces

Konstantinos Bampouras

TL;DR

This work investigates Nehari’s theorem for Hankel operators on Paley–Wiener spaces $PW(\Omega)$ under Schatten class constraints. It proves that Nehari’s theorem fails for all $p>4$ when $\Omega\subset\mathbb{R}^n$ is convex with $C^{2}$ boundary neighborhood of nonzero curvature, highlighting geometric obstructions to bounded symbols. For polytopes $P$, it establishes a Hardy-type inequality $\int_{2P}\frac{|\widehat{f}(x)|}{\omega_{P}(x)}dx\le C(P)\|f\|_{L^{1}}$ for $f\in PW^{1}(2P)$, which implies that any Hilbert–Schmidt Hankel operator on $PW(2P)$ is generated by a bounded symbol and yields Nehari for $p=2$ in the polytope case. The paper then introduces a generalized Hardy inequality with $\ rac{1}{\omega_{\Omega}^{d}}$ weights, identifying precise thresholds depending on geometry: $d\le 1$ for polytopes, $d<\frac{2}{n+1}$ (and fail for $d>1$) for smooth-boundary convex sets, and $d\le\frac{2}{n+1}$ in general, with sharp failures beyond these ranges. These results connect Schatten-class Hankel operator theory with geometric analysis of the ω-function and yield sharp bounds that inform when bounded-symbol Nehari-type characterizations can hold.

Abstract

Recently it was proven that for a convex subset of $\mathbb{R}^{n}$ that has infinitely many extreme vectors, the Nehari theorem fails, that is, there exists a bounded Hankel operator $\Ha_φ$ on the Paley--Wiener space $\PW(Ω)$ that does not admit a bounded symbol. In this paper we examine whether Nehari's theorem can hold under the stronger assumption that the Hankel operator $\Ha_φ$ is in the Schatten class $S^{p}(\PW(Ω))$. We prove that this fails for $p>4$ for any convex subset of $\mathbb{R}^{n}$, $n\geq2$, of boundary with a $C^{2}$ neighborhood of nonzero curvature. Furthermore we prove that for a polytope $P$ in $\mathbb{R}^{n}$, the inequality $$\int_{2P}\dfrac{|\widehat{f}(x)|}{m(P\cap (x-P))}dx\leq C(P)\|f\|_{L^{1}},$$ holds for all $f\in \PW^{1}(2P)$, and consequently any Hilbert--Schmidt Hankel operator on a Paley--Wiener space of a polytope is generated by a bounded function.

Nehari's Theorem and Hardy's inequality for Paley--Wiener spaces

TL;DR

This work investigates Nehari’s theorem for Hankel operators on Paley–Wiener spaces under Schatten class constraints. It proves that Nehari’s theorem fails for all when is convex with boundary neighborhood of nonzero curvature, highlighting geometric obstructions to bounded symbols. For polytopes , it establishes a Hardy-type inequality for , which implies that any Hilbert–Schmidt Hankel operator on is generated by a bounded symbol and yields Nehari for in the polytope case. The paper then introduces a generalized Hardy inequality with weights, identifying precise thresholds depending on geometry: for polytopes, (and fail for ) for smooth-boundary convex sets, and in general, with sharp failures beyond these ranges. These results connect Schatten-class Hankel operator theory with geometric analysis of the ω-function and yield sharp bounds that inform when bounded-symbol Nehari-type characterizations can hold.

Abstract

Recently it was proven that for a convex subset of that has infinitely many extreme vectors, the Nehari theorem fails, that is, there exists a bounded Hankel operator on the Paley--Wiener space that does not admit a bounded symbol. In this paper we examine whether Nehari's theorem can hold under the stronger assumption that the Hankel operator is in the Schatten class . We prove that this fails for for any convex subset of , , of boundary with a neighborhood of nonzero curvature. Furthermore we prove that for a polytope in , the inequality holds for all , and consequently any Hilbert--Schmidt Hankel operator on a Paley--Wiener space of a polytope is generated by a bounded function.

Paper Structure

This paper contains 4 sections, 23 theorems, 93 equations, 2 figures.

Key Result

Theorem 1.1

Suppose that $\Omega$ is a convex subset of $\mathbb{R}^{n}$, $n\geq 2$ that has a $C^{2}$ boundary neighborhood with nonzero curvature. Then Nehari's theorem fails for $\Omega$ for all $p>4$.

Figures (2)

  • Figure 1: Some useful geometric observations.
  • Figure 2: If $\Omega$ is the ellipse $\frac{x^2}{\gamma^2}+\frac{y^2}{\beta^2}<1$ then the vector $(0,\beta)$ belongs to the class $\mathscr{R}(\alpha,\beta,\beta)$.

Theorems & Definitions (43)

  • Theorem 1.1
  • Theorem 1.2
  • Theorem 1.3
  • Lemma 2.1
  • proof
  • Lemma 2.2
  • proof
  • Lemma 2.3
  • proof
  • Lemma 2.4
  • ...and 33 more