The Fourier transform in variable exponent Lebesgue spaces

André Kowacs, Wagner Augusto Almeida de Moraes · 2025

Read The Fourier transform in variable exponent Lebesgue spaces by André Kowacs, Wagner Augusto Almeida de Moraes (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-08891.txt.

Introduction

André Kowacs and Wagner de Moraes set a wave-shift form for fields whose power law shifts by place. The work dates to 2025. Such fields live in spaces where the growth rate is not one fixed count. The shift was not yet set for a wide class of them.

They set the shift as the slope of a smooth curve form for each field in the class. A size rule on the shift space makes it match the field space in full, length for length. They prove turn-back in that size and a trade rule for the shift. The work grounds new tools for this field kind.

Catalog record

CREATORAndré Kowacs, Wagner Augusto Almeida de Moraes
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE58 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-08891

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

The Fourier transform in variable exponent Lebesgue spaces André Kowacs, Wagner Augusto Almeida de Moraes Abstract . In this work we define a Fourier transform for each f ∈ L p ⁡ ( ⋅ ) ​ ( ℝ ) f\in L^{p(\cdot)}(\R) , for a large class of exponent functions p ⁡ ( ⋅ ) p(\cdot) , as the distributional derivative of a Hölder continuous function. A norm is defined in the space of such Fourier transforms so that it is isometrically isomorphic to L p ⁡ ( ⋅ ) ​ ( ℝ ) L^{p(\cdot)}(\R) . We also prove several properties of this Fourier transform, such as inversion in norm and an exchange theorem. 1991 Mathematics Subject Classification Primary: 42A38, 46E30. Secondary: 26A42, 46B04 1. Introduction The variable exponent Lebesgue spaces have been an area of intensive study in the last 30 years, with historical roots tracing back to the work of Orlicz in the 1930s on modular spaces [ 10 ] , which laid the groundwork for generalizations beyond constant exponents. The modern formulation of these spaces, particularly the variable exponent Lebesgue spaces L^{p(\cdot)}(\Omega) , where \Omega\subset\R^{n} , was significantly advanced by Kováčik and Rákosník [ 9 ] and Cruz-Uribe and Fiorenza [ 3 ] . These spaces can be seen as a natural generalization of the classical Lebesgue spaces. Roughly speaking, the usual constant parameter p\in[1,\infty] in the definition of L^{p}({\Omega}) is replaced by a measurable function p(\cdot):\mathbb{R}^{n}\to[1,\infty] . Although this generalization retains many properties of the classical Lebesgue spaces, the variable exponent Lebesgue spaces also display some very different phenomena depending on the exponent function p(\cdot) . Therefore extending results from L^{p} to L^{p(\cdot)} is in general a highly non-trivial problem and involves different techniques, a fact which has served as motivation for several papers in the last decades (see for instance [ 1 , 4 , 5 , 6 , 13 ] , and references therein). With this motivation, in the present work we extend several results from Talvila [ 12 ] . In that paper, the author introduced an alternative definition of the Fourier transform for functions f\in L^{p}(\mathbb{R}) , 1\leq p<\infty , establishing that the classical Fourier transform (seen as a tempered distribution) coincides with the distributional derivative of a Hölder continuous function. The work also develops additional properties of the Fourier transform \widehat{f} via the theory of the continuous primitive integral, including exchange theorems and norm inversion results. It is worth mentioning that when restricted to constant exponents, our results coincide precisely with the ones obtained in [ 12 ] . Moreover, while many papers on variable exponent Lebesgue spaces restrict to the case where the exponent function p(\cdot) assumes values in [1,\infty) , in this paper we consider the full range of values [1,\infty] , which allows us to obtain more general results. Notably, when the essential supremum of p(\cdot) is infinite, the analysis diverges significantly from the constant exponent case (see, for example, [ 8 ] ). This is due to some characteristics of the structure of L^{\infty}(\mathbb{R}) , such as the lack of certain approximation properties and the failure of density results, requiring distinct techniques compared to finite exponent scenarios. The paper is structured as follows: Section 2 reviews the fundamental theory of variable exponent Lebesgue spaces, and fix the notation that will be used throughout the paper. Section 3 introduces a key integral kernel function for Fourier analysis, establishing its regularity under diverse exponent conditions through Hölder and Lipschitz continuity estimates. In Section 4 we define two Banach spaces, while describing the Fourier transform rigorously as a distributional derivative, and show that these spaces are isometrically isomorphic to L^{p(\cdot)}(\mathbb{R}) . Section 5 derives operator calculus for this transform, establishing its consistency with tempered distributions and invariance under translation operations. Finally, Section 6 develops an integration framework for Fourier transforms, proving both exchange theorems and norm-convergent inversion formulas that extend classical Fourier analysis. 2. Preliminaries For the main theory and results listed in this section, we refer to [ 3 ] . Throughout this paper, we fix the usual Lebesgue measure on the real line. Given a set \Omega\subset\R^{n} , let \mathcal{P}(\Omega) denote the set of all measurable functions p:\Omega\to[1,\infty] . These are called exponent functions, and to distinguish between variable and constant exponents, we will denote them by p(\cdot) . We also fix the following notations: \Omega_{1}^{p(\cdot)}=\{x\in\Omega:p(x)=1\}, \Omega_{*}^{p(\cdot)}=\{x\in\Omega:1<p(x)<\infty\}, \Omega_{\infty}^{p(\cdot)}=\{x\in\Omega:p(x)=\infty\}, p_{-}=\underset{{x\in\Omega}}{\text{ess inf}}\{p(x)\}, p_{+}=\underset{{x\in\Omega}}{\text{ess sup}}\{p(x)\}. Given p(\cdot)\in\mathcal{P}(\Omega) , its conjugate exponent q(\cdot)\in\mathcal{P}(\Omega) is defined by \frac{1}{p(x)}+\frac{1}{q(x)}=1,\quad x\in\Omega. In this paper, q(\cdot) will always denote the respective conjugate exponent of any given p(\cdot) . Note that as a consequence of these definitions, the following symmetric identities always hold: \frac{1}{p_{+}}+\frac{1}{q_{-}}=1, \frac{1}{p_{-}}+\frac{1}{q_{+}}=1. A key condition for the exponent functions which ensures important properties for the L^{p(\cdot)}(\Omega) spaces (see [ 14 , 15 , 7 ] ), is the so called log-Hölder continuity. An exponent p(\cdot)\in\mathcal{P}(\Omega) is locally log-Hölder continuous if there exists a constant c>0 such that |p(x)-p(y)|\leq\frac{c}{\log(e+1/|x-y|)}, for every x,y\in\Omega . Also, p(\cdot) is log-Hölder continuous at infinity if there exist constants C,p_{\infty}>0 such that (1) |p(x)-p_{\infty}|\leq\frac{C}{\log(e+|x|)}, for every x\in\Omega . If \frac{1}{p(\cdot)} is log-Hölder continuous at infinity, we say that it satisfies the log-Hölder decay condition. In this paper we will assume only that ( 1 ) holds for almost every x\in\R . Given p(\cdot)\in\mathcal{P}(\Omega) and \Omega^{\prime}\subset\Omega the modular \rho_{p(\cdot),\Omega^{\prime}} associated with p(\cdot) denotes the functional defined on Lebesgue measurable functions given by \rho_{p(\cdot),\Omega^{\prime}}(f)=\int_{\Omega^{\prime}\backslash\Omega_{\infty}^{p(\cdot)}}|f(x)|^{p(x)}dx+\|f\|_{L^{\infty}(\Omega^{\prime}\cap\Omega_{\infty}^{p(\cdot)})}\in[0,\infty]. Often during this work \Omega^{\prime}=\Omega , and we abbreviate the modular by \rho_{p(\cdot)}(f) . The variable exponent Lebesgue space L^{p(\cdot)}(\Omega) is then defined as the set of measurable functions f such that \rho_{p(\cdot)}(f/\lambda)<\infty , for some \lambda>0 , with the usual convention that two functions are “equal” if they coincide almost everywhere. The norm in L^{p(\cdot)} is then defined as follows. Given \Omega\subset\R^{n} and p(\cdot)\in\mathcal{P}(\Omega) , define \|f\|_{L^{p(\cdot)}(\Omega)}=\inf\{\lambda>0:\rho_{p(\cdot),\Omega}(f/\lambda)\leq 1\}, for every Lebesgue measurable function f , with the standard convention that \inf\varnothing=\infty . When the set \Omega is implicit, we will often abbreviate the norm by \|f\|_{p(\cdot)} . Remark 1 . Notice that \|f\|_{L^{p(\cdot)}(\Omega)}<\infty for every f\in L^{p(\cdot)}(\Omega) . In particular, if \rho_{p(\cdot)}(f)<\infty , then \|f\|_{L^{p(\cdot)}(\Omega)}<\infty as well. The variable exponent Lebesgue spaces extend the usual Lebesgue spaces, and share similar properties. In particular, the variable exponent Lebesgue space norm satisfies a generalization of Hölder’s inequality, presented below. First, we introduce a notation to simplify several statements throughout this paper. Definition 2.1 . For a Lebesgue measurable set E\subset\R^{n} , denote by |E| its Lebesgue measure. Define \delta(E)=\|\chi_{E}\|_{L^{\infty}(\R^{n})} , that is, \delta(E)=0 if |E|=0 and \delta(E)=1 , otherwise. Theorem 2.2 (Generalized Hölder’s Inequality) . Given \Omega\subset\R^{n} and p(\cdot)\in\mathcal{P}(\Omega) , for all f\in L^{p(\cdot)}(\Omega) and g\in L^{q(\cdot)}(\Omega) , fg\in L^{1}(\Omega) and \int_{\Omega}|f(x)g(x)|dx\leq K_{p(\cdot)}\|f\|_{p(\cdot)}\|g\|_{q(\cdot)}, where K_{p(\cdot)}=\left(\frac{1}{p_{-}}-\frac{1}{p_{+}}+1\right)\delta\left(\Omega_{*}^{p(\cdot)}\right)+\delta\left(\Omega_{\infty}^{p(\cdot)}\right)+\delta\left(\Omega_{1}^{p(\cdot)}\right), and q(\cdot)\in\mathcal{P}(\Omega) is the conjugate exponent of p(\cdot) . 3. Properties of \Psi_{f} In this section we define, for each f\in L^{p(\cdot)}(\R) , an auxiliary function \Psi_{f}:\R\to\R which will be essential in defining the Fourier transform of f . We also prove several properties of this function \Psi_{f} , based on properties of the exponent p(\cdot) . Definition 3.1 . Let f\in L^{p(\cdot)}(\R) , for some p(\cdot)\in\mathcal{P}(\R) . Define \Psi_{f}(s)\vcentcolon=\int_{-\infty}^{\infty}\left(\frac{1-e^{-ist}}{it}\right)f(t)dt, for s\in\R\backslash\{0\} , whenever the integral is well defined, and \Psi_{f}(0)=0 . For 1<q<\infty ,

Read the complete text