Table of Contents
Fetching ...

How smooth are restrictions of Besov functions?

Julien Brasseur

TL;DR

This work resolves a sharpness question for the restriction of Besov functions: while Besov spaces B_{p,q}^s generally fail to restrict to a fixed dimension when p<q, their partial maps can be controlled by Besov spaces of generalised smoothness B_{p,p}^{(s,Ψ)} with a slowly varying modulation Ψ, provided a simple summability condition holds. The authors prove that this condition is also necessary, by constructing f ∈ B_{p,q}^s(R^N) whose partial maps f(·,y) fail to lie in B_{p,∞}^{(s,Ψ)}(R^d) for almost every y whenever Ψ violates the summability criterion. The main technical tool is a carefully crafted sequence (λ_{j,k}) whose global Besov norm remains finite, but whose dyadic-locally weighted sums diverge along every dyadic line when weighted by Ψ, enabled by a rearrangement argument. The result completes the picture by providing a complete, sharp dichotomy for the restriction behavior of Besov functions in terms of generalized smoothness and slowly varying modulations, clarifying the role of the parameter q relative to p and the impact of Ψ on restriction properties.

Abstract

In a previous work, we showed that Besov spaces do not enjoy the restriction property unless $q\leq p$. Specifically, we proved that if $p<q$, then it is always possible to construct a function $f\in B_{p,q}^s(\mathbb{R}^N)$ such that $f(\cdot,y)\notin B_{p,q}^s(\mathbb{R}^d)$ for a.e. $y\in \mathbb{R}^{N-d}$, while this "pathology" does not happen if $q\leq p$. We showed that the partial maps belong, in fact, to the Besov space of generalised smoothness $B_{p,q}^{(s,Ψ)}(\mathbb{R}^d)$ provided the function $Ψ$ satisfies a simple summability condition involving $p$ and $q$. This short note completes the picture by showing that this characterisation is sharp.

How smooth are restrictions of Besov functions?

TL;DR

This work resolves a sharpness question for the restriction of Besov functions: while Besov spaces B_{p,q}^s generally fail to restrict to a fixed dimension when p<q, their partial maps can be controlled by Besov spaces of generalised smoothness B_{p,p}^{(s,Ψ)} with a slowly varying modulation Ψ, provided a simple summability condition holds. The authors prove that this condition is also necessary, by constructing f ∈ B_{p,q}^s(R^N) whose partial maps f(·,y) fail to lie in B_{p,∞}^{(s,Ψ)}(R^d) for almost every y whenever Ψ violates the summability criterion. The main technical tool is a carefully crafted sequence (λ_{j,k}) whose global Besov norm remains finite, but whose dyadic-locally weighted sums diverge along every dyadic line when weighted by Ψ, enabled by a rearrangement argument. The result completes the picture by providing a complete, sharp dichotomy for the restriction behavior of Besov functions in terms of generalized smoothness and slowly varying modulations, clarifying the role of the parameter q relative to p and the impact of Ψ on restriction properties.

Abstract

In a previous work, we showed that Besov spaces do not enjoy the restriction property unless . Specifically, we proved that if , then it is always possible to construct a function such that for a.e. , while this "pathology" does not happen if . We showed that the partial maps belong, in fact, to the Besov space of generalised smoothness provided the function satisfies a simple summability condition involving and . This short note completes the picture by showing that this characterisation is sharp.

Paper Structure

This paper contains 4 sections, 5 theorems, 31 equations, 1 figure.

Key Result

Theorem 1.1

Let $N\geqslant2$, $1\leqslant d<N$, $0<p<q\leqslant\infty$, $s>\sigma_p$ and let $\Psi$ be a slowly varying function satisfying cond:psi. Suppose that $f\in B_{p,q}^s(\mathbb{R}^N)$. Then,

Figures (1)

  • Figure 1: Each rectangle represents a term of the series $\sum u_j/U_j^m$ and the continuous line the curve $1/x^m$. The gray rectangle has width $U_j-U_{j-1}=u_j$ and height $1/U_j^m$.

Theorems & Definitions (18)

  • Theorem 1.1
  • Remark 1.2
  • Remark 1.2
  • Theorem 1.3
  • Remark 1.4
  • Remark 1.5
  • Definition 2.1
  • Remark 2.2
  • Definition 2.3
  • Remark 2.4
  • ...and 8 more