Regularity of fractional maximal functions through Fourier multipliers

DOI10.1016/J.JFA.2018.11.004MaRDI QIDQ1715469zbMATH OpenOpenAlexWikidataFDO

Authors David Beltran, João Pedro Ramos, Olli Saari

Publication date 4 February 2019

Published in Journal of Functional Analysis (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1803.02581

bounded variation maximal function Sobolev space spherical mean, fractional maximal function

Multipliers for harmonic analysis in several variables (42B15) Maximal functions, Littlewood-Paley theory (42B25) Sobolev spaces and other spaces of ``smooth functions, embedding theorems, trace theorems (46E35)

Abstract: We prove endpoint bounds for derivatives of fractional maximal functions with either smooth convolution kernel or lacunary set of radii in dimensions

n g e q 2

. We also show that the spherical fractional maximal function maps

L^{p}

into a first order Sobolev space in dimensions

n g e q 5

.

Review text

The paper under review studies regularity properties of maximal functions in the fractional setting. More precisely the authors consider the operators \[ M_\alpha ^{lac}f(x):= \sup_{k\in\mathbb{Z}} \Big|\frac{2^{\alpha k}}{|B(0,2^k)|}\int_{B(x,2^k)}f(y) dy\Big|,\quad M_\alpha ^{\phi}(x) :=\sup_{t>0}t^\alpha |f*\phi_t(x)|,\quad x\in\mathbb{R}^n, \] where \(\alpha\in[0,n)\) in general and \(\phi\) is a positive Schwartz function and \(\phi_t(x):=t^{-n}\phi(x/t)\). The objective is to study suitable mapping properties of these operators on spaces involving derivatives and is motivated by a series of studies and previous results concerning the action of maximal operators on Sobolev spaces. The main result of the paper is the statement that if \(\mathcal{M}_\alpha\in\{M_\alpha ^{lac},M_\alpha ^{\phi}\}\), \(\alpha\in(0,1)\), \(n\geq 2\) and \(p=n/(n-\alpha)\) then \[ \|\nabla\mathcal{M}_\alpha f\|_{L^p(\mathbb{R}^n)}\lesssim |f| _{\mathrm{BV}(\mathbb{R}^n)}. \] For \(\alpha\geq 1\) the estimate is still valid but it is a relatively easy consequence of a pointwise inequality of \textit{J. Kinnunen} and \textit{E. Saksman} [Bull. Lond. Math. Soc. 35, No. 4, 529--535 (2003; Zbl 1021.42009)], together with the Gagliardo-Nirenberg-Sobolev inequality (as observed in [\textit{E. Carneiro} and \textit{J. Madrid}, Trans. Am. Math. Soc. 369, No. 6, 4063--4092 (2017; Zbl 1370.26022)]). In one dimension the corresponding result for the non-centered fractional maximal function is in [Carneiro and Madrid, loc. cit.]. The case \(\alpha=0\) corresponds to the usual (centered) maximal function, or its smooth variant and the relevant inequality is a known (and hard) open problem in dimensions \(n\geq 2\). One of the novelties of the paper is in the use of Fourier analytic techniques in the context of regularity questions for maximal functions. Roughly speaking, the proof proceeds by showing a single scale estimate (one where the supremum is over radii \(1\leq t\leq 2\)) in the scale of Besov spaces. The proof uses multiplier theorems for the corresponding symbols, interpolation, Bernstein's inequality, and other classical estimates. The estimate can be extended to the full supremum when the symbols involved are nice (e.g. smooth kernels, or dilations corresponding to lacunary radii). As the authors mention these proofs above are to some extent inspired by the proof of Stein's spherical maximal theorem. The authors also provide a regularity estimate for the fractional spherical maximal function for dimensions \(n\geq 5\) of the form \[ \|\nabla S_\alpha f\|_{L^q}\lesssim \|f\|_{L^p},\quad \frac{n}{n-2}< p\leq q<\infty, \quad \frac1q=\frac1p-\frac{\alpha-1}{n},\quad 1\leq \alpha<\alpha(p), \] and suitable \(\alpha=\alpha(p)\). The proof of this result is also Fourier analytic.

Recommendations

Cites work

Cited in

(24)

This page was built for publication: Regularity of fractional maximal functions through Fourier multipliers

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1715469)