Composition operators on weighted modulation spaces
Héctor Ariza, Carmen Fernández, Antonio Galbis · 2025
Read Composition operators on weighted modulation spaces by Héctor Ariza, Carmen Fernández, Antonio Galbis (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-2512-15409.txt.
Introduction
Hector Ariza, Carmen Fernandez, and Antonio Galbis study composition operators on weighted modulation spaces in a 2025 math paper. A composition operator reshapes a function by feeding its input through another function first. Modulation spaces measure a function by both its size and its frequency content, with weights that stress some parts over others.
The authors look at operators whose symbols are suitable small changes of the identity map. They study how these operators act between different weighted modulation classes. The work covers spaces built from tempered distributions and also spaces of ultradistributions defined with a subadditive weight. The results clarify when such reshaping stays well behaved across these settings.
Catalog record
| CREATOR | Héctor Ariza, Carmen Fernández, Antonio Galbis |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 51 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-15409
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Composition operators on weighted modulation spaces Héctor Ariza, Carmen Fernández, Antonio Galbis Abstract We study composition operators whose symbols are suitable perturbations of the identity and which act between different weighted modulation classes. We consider both modulation spaces formed by tempered distributions and those whose elements are ultradistributions defined in terms of a subadditive weight. Mathematics Subject Classification (2010): Primary 47G30; Secondary 42B35, 46F05, 47B33. Keywords: Composition operators, pseudodifferential operators, modulation spaces, Gelfand-Shilov spaces. Emails: [email protected], [email protected], [email protected] The research of C. Fernández and A. Galbis is supported by project PID2024-162128NB-I00, Agencia Estatal de Investigación, Ministerio de Ciencia, Investigación y Universidades. Composition operators on various spaces of smooth functions have been extensively studied during the past ten years. A central problem is to find necessary and/or sufficient conditions under which the operator on a given space is well defined and continuous. For the Schwartz space {\mathcal{S}}({\mathbb{R}}) this was settled in [ 17 ] . In a recent paper [ 4 ] the authors consider weighted composition operators in {\mathcal{S}}({\mathbb{R}}^{d}). The study of composition operators in Gelfand-Shilov spaces {\mathcal{S}}_{\omega}({\mathbb{R}}) began in [ 2 ] and continued in [ 3 , 1 ] . The spaces {\mathcal{S}}_{\omega}({\mathbb{R}}) are versions of the Schwartz class in the ultradifferentiable setting, the most important examples being the classical Gelfand-Shilov spaces \Sigma_{s}, s>1. Both the Schwartz class and Gelfand-Shilov spaces can be obtained as intersections of weighted modulation and ultra-modulation spaces. It is therefore natural to consider composition operators on weighted (ultra-)modulation spaces. In 1953, Beurling and Helson proved that if \psi is continuous on the torus and the corresponding composition operator is continuous on the algebra of absolutely convergent Fourier series, then \psi(t)=kt+\psi(0) for some integer k. Later, Lebedev and Olevskii [ 21 ] showed that if \psi:{\mathbb{R}}^{d}\to{\mathbb{R}}^{d} is {\mathcal{C}}^{1} and defines a continuous composition operator from {\mathcal{F}}L^{p}({\mathbb{R}}^{d}) into itself for 1\leq p<\infty , p\neq 2, then \psi must be an affine bijection. More recently, Oukoudjou [ 24 ] (see also [ 26 ] ) extended this result to modulation spaces M^{p,q}({\mathbb{R}}^{d}) under the assumption 2\neq q<\infty. Our aim in this paper is to establish sufficient conditions on a smooth function \psi:{\mathbb{R}}^{d}\to{\mathbb{R}}^{d} to ensure that the composition operator is well-defined and continuous between two weighted (ultra-)modulation spaces. Using the Fourier inversion formula, one has f(\psi(x))=\int_{{\mathbb{R}}^{d}}\widehat{f}(\xi)e^{2\pi i\psi(x)\xi}d\xi,\,f\in{\mathcal{S}}({\mathbb{R}}^{d}), which suggest studying Fourier integral operators with symbol identically equal to 1 and phase \Phi(x,\xi)=\psi(x)\xi. Fourier integral operators on modulation spaces have been investigated in a series of papers (see [ 10 , 16 ] and the references therein). The conditions imposed on the phase imply that: 1. |\partial_{x}^{\alpha}\psi(x)|\leq C_{\alpha}, |\alpha|\geq 1,x\in{\mathbb{R}}^{d}. 2. |\mbox{det}\left(\frac{\partial\psi_{j}}{\partial x_{i}}(x)\right)|\geq\delta>0, x\in{\mathbb{R}}^{d}. The second condition, known as the non-degeneracy condition, imposes significant restrictions for our purposes. Instead, we represent the composition operator as a Kohn-Nirenberg pseudodifferential operator \begin{array}[]{*2{>{\displaystyle}l}}f\left(\psi(x)\right)&=\int_{{\mathbb{R}}^{d}}\widehat{f}(y)e^{2\pi iy\psi(x)}\,dy\\ &\\ &=\int_{{\mathbb{R}}^{d}}\sigma(x,y)\widehat{f}(y)e^{2\pi ixy}\,dy=\sigma(x,D)f(x),\end{array} with symbol \sigma(x,y)=\exp\left(i\phi(x)y\right),\ \phi(x)=2\pi(\psi(x)-x). We thus study pseudodifferential operators whose symbols have the form \sigma(x,y)=\exp\left(i\phi(x)y\right) under suitable assumptions on \phi. As a consequence, we obtain continuity results for certain composition operators acting between different (ultra-) modulation classes. In the case d=1, it is worth noting that [ 2 ] showed that composition with a polynomial of degree greater than one never leaves Gelfand-Shilov spaces invariant; however, positive results were obtained when \psi satisfies |x|\leq C(1+|\psi(x)|) and \psi^{\prime}\in{\mathcal{B}}_{L_{\infty},\omega} (defined below, see [ 2 , Corollary 4.2] ). Moreover, by [ 2 , Theorem 3.9] , if the composition with \psi leaves \Sigma_{s}({\mathbb{R}}) invariant, then \psi^{\prime} must be bounded. This motivates the type of conditions we impose on \phi. The paper is organized as follows. In 2 , we study continuity results for certain composition operators on weighted modulation spaces M^{p}_{m}({\mathbb{R}}^{d}) for 1\leq p\leq\infty. Section 3 considers the analogous problem for ultramodulation spaces associated with subadditive weight functions in the sense of Braun, Meise and Taylor, in the one dimensional setting; the results are then extended to several variables. 1 Preliminaries 1.1 Classes of ultradifferentiable functions Definition 1.1 . A continuous increasing function \omega:[0,\infty[\longrightarrow[0,\infty[ is called a subadditive weight , or shortly a weight, if it satisfies: (\alpha) \omega(s+t)\leq\omega(s)+\omega(t) for all s,\,t\geq 0 , (\beta) \displaystyle\int_{0}^{\infty}\frac{\omega(t)}{1+t^{2}}\ dt<\infty , (\gamma) \log(1+t^{2})=o(\omega(t)) as t tends to \infty , (\delta) \varphi_{\omega}:t\rightarrow\omega(e^{t}) is convex. \omega is extended to {\mathbb{R}}^{d} as \omega(x)=\omega(|x|). The Young conjugate \varphi_{\omega}^{*}:[0,\infty[\longrightarrow{\mathbb{R}} of \varphi_{\omega} is defined by \varphi_{\omega}^{*}(s):=\sup\{st-\varphi_{\omega}(t):\ t\geq 0\},\ s\geq 0. The following inequality will be used throughout the paper: 3m\varphi_{\omega}^{\ast}(\frac{j}{3m})+j\leq m\varphi_{\omega}^{\ast}(\frac{j}{m})\mbox{ for all }j,\,m\in{\mathbb{N}}. (1.1) In fact, since \varphi_{\omega}(x+1)=\omega(ee^{x})\leq 3\omega(e^{x})=3\varphi_{\omega}(x), we have \begin{array}[]{c}3\varphi_{\omega}^{\ast}(\frac{j}{3m})+\frac{j}{m}=\sup_{x\geq 0}\{(x+1)\frac{j}{m}-3\varphi_{\omega}(x)\}\\ \\ \leq\sup_{x\geq 0}\{(x+1)\frac{j}{m}-\varphi_{\omega}(x+1)\}\leq\varphi_{\omega}^{\ast}(\frac{j}{m}).\end{array} Definition 1.2 . For a weight function \omega we define the \omega -ultradifferentiable functions on {\mathbb{R}}^{d} of Beurling type, \mathcal{E}_{\omega}({\mathbb{R}}^{d}) , as \begin{split}\mathcal{E}_{\omega}({\mathbb{R}}^{d})=\{f\in C^{\infty}({\mathbb{R}}^{d}):\textit{ for each compact }K\textit{ and for each }m\in\mathbb{N}:\\ p_{K,m}(f)<\infty\}\end{split} where p_{K,m}(f)=\sup_{x\in K,\alpha\in\mathbb{N}_{0}^{d}}|f^{(\alpha)}(x)|\exp\left(-m\varphi^{*}_{\omega}(\frac{|\alpha|}{m})\right) . \mathcal{E}_{\omega}({\mathbb{R}}^{d}) endowed with the topology induced by the seminorms p_{K,m} m\in{\mathbb{N}} and K running in the compact sets of {\mathbb{R}}^{d} is a Fréchet space. Condition ( \beta ) for the weight implies that \mathcal{D}_{\omega}({\mathbb{R}}^{d}):=\mathcal{E}_{\omega}({\mathbb{R}}^{d})\cap\mathcal{D}({\mathbb{R}}^{d}) is non trivial. The spaces \mathcal{B}_{L_{\infty},\omega}({\mathbb{R}}^{d}) were defined in [ 5 ] as extensions of the classical space \mathcal{B}_{L_{\infty}}({\mathbb{R}}^{d}), introduced by Schwartz, of bounded smooth functions having bounded derivatives. \mathcal{B}_{L_{\infty},\omega}({\mathbb{R}}^{d}) consists of those functions f\in C^{\infty}({\mathbb{R}}^{d}) such that for every m\in\mathbb{N} there exists C_{m} satisfying \left|f^{(\alpha)}(x)\right|\leq C_{m}\hskip 1.42271pt\exp\left(m\varphi_{\omega}^{*}(\frac{|\alpha|}{m})\right) (1.2) for all \alpha\in\mathbb{N}_{0}^{d} and x\in\mathbb{R}^{d} . It is endowed with the topology defined by the sequence of seminorms \sup_{x\in{\mathbb{R}}^{d}}\sup_{\alpha\in{\mathbb{N}}_{0}^{d}}\left|f^{(\alpha)}(x)\right|\exp\left(-m\varphi_{\omega}^{*}(\frac{|\alpha|}{m})\right),\,m\in{\mathbb{N}}. The convexity of \varphi^{\ast}_{\omega} implies that we can replace m\varphi_{\omega}^{*}(\frac{|\alpha|}{m}) by m\varphi_{\omega}^{*}(\frac{|\alpha|+1}{m}) in the definition of \mathcal{B}_{L_{\infty},\omega}({\mathbb{R}}^{d}) as well as in the seminorms that define its topology. For a space of functions E(\mathbb{R}^{d}), we put E(\mathbb{R}^{d},\mathbb{R}^{d})=\left\{(\phi_{1},\dots,\phi_{d}):\ \phi_{i}\in E(\mathbb{R}^{d})\right\}. The spaces \mathcal{S}_{\omega}(\mathbb{R}^{d}) were defined by Björck. We give an equivalent definition [ 6 ] . Definition 1.3 . Let \omega be a weight function. The Gelfand-Shilov space of Beurling type \mathcal{S}_{\omega}(\mathbb{R}^{d}) consists of those functions f\in C^{\infty}({\mathbb{R}}^{d}) such that \pi_{\lambda,\mu}(f):=\sup_{\alpha\in{\mathbb{N}}_{0}^{d}}\sup_{x\in{\mathbb{R}}^{d}}|f^{(\alpha)}(x)|e^{-\lambda\varphi^{\ast}(\frac{|\alpha|}{\lambda})+\mu\omega(x)}<\infty,\ \ \mbox{for all }\lambda>0,\mu>0. The classical Gelfand-Shilov spaces \Sigma_{s}({\mathbb{R}}^{d}), s>1, are defined by the weights \omega(t)=t^{1/s}. 1.2 Time-frequency analysis Given f,\,g\in L^{2}({\mathbb{R}}^{d}), g\neq 0, the short time Fourier transform (STFT) of f with respect to the window g is defined as V_{g}f(x,\omega)=\displaystyle\int_{{\mathbb{R}}^{d}}f(t)\overline{g(t-x)}e^{-2\pi it\omega}\,dt. The STFT satisfies the ortogonality relations [ 18 , 3.2.1] \langle V_{g_{1}}f_{1},V_{g_{2}}f_{2}\rangle=\langle f_{1},f_{2}\rangle\overline{\langle g_{1},g_{2}\rangle}, (1.3) where the brackets stand for the inner product in L^{2}({\mathbb{R}}^{d}). We will also use brackets \langle f,g\rangle to denote the extension to {\mathcal{S}}^{\prime}({\mathbb{R}}^{d})\times{\mathcal{S}}({\mathbb{R}}^{d}) (or to {\mathcal{S}}_{\omega}^{\prime}({\mathbb{R}}^{d})\times{\mathcal{S}}_{\omega}({\mathbb{R}}^{d}) ) of the inner product in L^{2}({\mathbb{R}}^{d}). The modulation and translation operators are defined by M_{\omega}f(t)=e^{2\pi\imath\omega t}f(t) and T_{x}f(t)=f(t-x). For a non-zero g\in{\mathcal{S}}({\mathbb{R}}^{d}) (resp. g\in{\mathcal{S}}_{\omega}({\mathbb{R}}^{d}) ) and a tempered distribution f\in{\mathcal{S}}^{\prime}({\mathbb{R}}^{d}) (resp. f\in{\mathcal{S}}_{\omega}^{\prime}({\mathbb{R}}^{d}) ), the STFT of f with window g is given by the continuous function V_{g}f(x,\omega)=\left<f,M_{\omega}T_{x}g\right>. A positive function v on {\mathbb{R}}^{N}, is said to be submultiplicative if v(x_{1}+x_{2})\leq v(x_{1})v(x_{2}),\ \ x_{1},x_{2}\in{\mathbb{R}}^{N}. If v is submultiplicative, a positive function m is called v -moderate if for some constant C>0 , m(x_{1}+x_{2})\leq Cm(x_{1})v(x_{2}),\ \ x_{1},x_{2}\in{\mathbb{R}}^{N}. Important examples of submultiplicative weights are the functions (1+|z|)^{s}\ \ (s\geq 0),\ v(z)=e^{a|z|},\ e^{a\omega(z)}\ \ (a>0), when \omega is a weight function. The weights m(z)=(1+|z|)^{s} (s\in{\mathbb{R}}) are (1+|z|)^{|s|} -moderate and the weights e^{a\omega(z)}, ( a\in{\mathbb{R}} ), are e^{|a|\omega(z)} -moderate. Let m be a (1+|z|)^{s} -moderate weight on {\mathbb{R}}^{2d} for some s>0 . We denote by M_{m}^{p}({\mathbb{R}}^{d}) the space M_{m}^{p}({\mathbb{R}}^{d}):=\{f\in{\mathcal{S}}^{\prime}({\mathbb{R}}^{d}):\,m\,V_{g}f\in L^{p}({\mathbb{R}}^{2d})\} endowed with the norm \|f\|_{M_{m}^{p}}=\|m\,V_{g}f\|_{L^{p}}. If m\equiv 1 , we