An H 1 H^{1} multiplier theorem on anisotropic Hardy sapces
Yiyu Tang · 2025
Read An H 1 H^{1} multiplier theorem on anisotropic Hardy sapces by Yiyu Tang (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-2508-17355.txt.
Introduction
Yiyu Tang proves a multiplier theorem for anisotropic Hardy spaces in this short 2025 paper. The classical Sledd-Stegenga theorem says which multipliers carry the H1 space into L1; Tang carries that result into settings where different directions scale at different rates. With the same technique, the paper also proves an H1 to Lp version that holds for every p from one upward.
Catalog record
| CREATOR | Yiyu Tang |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 50 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-17355
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
An H 1 H^{1} multiplier theorem on anisotropic Hardy sapces Yiyu Tang Abstract . A parallel result of Sledd–Stegenga’s H 1 → L 1 H^{1}\rightarrow L^{1} multiplier theorem was obtained on the H 1 H^{1} space under the anisotropic settings. Based on the same technique, an H 1 → L p H^{1}\rightarrow L^{p} multiplier theorem is also proved for 1 ≤ p < ∞ 1\leq p<\infty . 1. Introduction Let H^{p}(\mathbb{R}^{d}) be the real Hardy spaces (we assume 0<p\leq 1 ). The following Hardy–Littlewood inequality is well known (see, for example, [ 9 ] , Chapter \RN 3, Section 5.4) \Big(\int_{\mathbb{R}^{d}}\frac{|\widehat{f}|^{p}}{|\xi|^{(2-p)d}}\,\mathrm{d}\xi\Big)^{\frac{1}{p}}\lesssim_{p,d}\|f\|_{H^{p}(\mathbb{R}^{d})}. When p=1 , we have a stronger result, due to Sledd and Stegenga (see [ 8 ] , Theorem 1). The H^{1} boundedness criterion. Let \mu be a positive Borel measure on \mathbb{R}^{d}\setminus\{0\} , then \int_{\mathbb{R}^{d}}|\widehat{f}(\xi)|\,\mathrm{d}\mu(\xi)\lesssim\|f\|_{H^{1}(\mathbb{R}^{d})}, if and only if \sup_{\epsilon>0}\Big(\sum_{\alpha\in\mathbb{Z}^{d}\setminus\{0\}}\mu\big(Q_{\epsilon}(\alpha)\big)^{2}\Big)^{\frac{1}{2}}<\infty, where Q_{\epsilon}(\alpha) is the cube \{(x_{i})_{i=1}^{d}\in\mathbb{R}^{d}:\epsilon\alpha_{i}-\frac{\epsilon}{2}\leq x_{i}<\epsilon\alpha_{i}+\frac{\epsilon}{2}\} . It is known that, under the anisotropic settings H^{p}_{A} (we assume 0<p\leq 1 ), the Hardy–Littlewood inequality still holds true: Let A be a dilation matrix, and \rho_{\ast}(\cdot) is a quasi-norm associated with A^{\ast} , then \Big(\int_{\mathbb{R}^{d}}\frac{|\widehat{f}|^{p}}{\rho_{\ast}(\xi)^{2-p}}\,\mathrm{d}\xi\Big)^{\frac{1}{p}}\lesssim_{p,d,A}\|f\|_{H^{p}_{A}(\mathbb{R}^{d})}. This inequality was proved by Bownik–Wang, see [ 3 ] , Corollary 8. However, the anisotropic analog of Sledd–Stegenga’s criterion seems to be absent so far. We will close this gap. The statement of the criterion under the anisotropic settings is more complicated and requires more preparation. Moreover, some arguments in [ 8 ] does not work any more, and we need few modifications. Also, there are problems in the original proofs of [ 8 ] , which either omit many non-trivial steps or are too specific (only the 1-dimensional case is proved). Here we complete the missing details of the proof in [ 8 ] . 2. Preliminaries In this section, most of the text is adapted from the book [ 2 ] . 2.1. Dilation matrix and its homogeneous norm Let A\in\mathbf{GL}(d,\mathbb{R}) and b\coloneqq|\det(A)| . We say that A is a dilation matrix , if all its eigenvalues \{\lambda_{i}\}_{i=1}^{d} have modulus strictly greater than 1 . For each dilation matrix A , there exists a measurable function (not unique) \rho_{A}\colon\mathbb{R}^{d}\rightarrow[0,\infty) , called a homogenous quasi-norm associated with A , satisfying the following properties: (1) \rho_{A}(x)=0 if and only if x=0 . (2) \rho_{A}(Ax)=b\rho_{A}(x) for all x\in\mathbb{R}^{d} . (3) \rho_{A}(x+y)\leq c(\rho_{A}(x)+\rho_{A}(y)) for all x,y\in\mathbb{R}^{d} , where c>0 is an absolute constant independent of x and y . We choose \lambda_{-} and \lambda_{+} so that 1<\lambda_{-}<|\lambda_{1}|\leq\ldots\leq|\lambda_{d}|<\lambda_{+} . The following growth estimate of \rho_{A} is known. Proposition 2.1 . (see [ 6 ] , Appendix B) Suppose that \rho_{A} is a homogenous quasi-norm associated with the matrix A , then there is a constant c_{A}>0 , so that \displaystyle c_{A}^{-1}\rho_{A}(x)^{\zeta_{-}} \displaystyle\leq|x|\leq c_{A}\rho_{A}(x)^{\zeta_{+}},\text{ if }\rho_{A}(x)\geq 1, \displaystyle c_{A}^{-1}\rho_{A}(x)^{\zeta_{+}} \displaystyle\leq|x|\leq c_{A}\rho_{A}(x)^{\zeta_{-}},\text{ if }\rho_{A}(x)<1. Here \zeta_{\pm}\coloneqq\ln\lambda_{\pm}/\ln b , and c_{A} only depends on \zeta_{\pm} . The basic geometric objects in isotropic settings are Euclidean balls B(x,r) . For the anisotropic case, we have the following substitute. We say that \Delta\subset\mathbb{R}^{d} is an ellipse , if there exists Q\in\mathbf{GL}(d,\mathbb{R}) , so that \Delta=Q^{-1}(B(0,1)) , where B(0,1) is the (open) unit Euclidean ball in \mathbb{R}^{d} . Proposition 2.2 . (See [ 2 ] , Chapter 1, Lemma 2.2) Suppose that A is a dilation matrix, then there exists an ellipse \Delta\subset\mathbb{R}^{d} , so that \Delta\subset A(\Delta). This proposition is similar to the fact that B(0,r)\subset B(0,R) whenever r<R . By rescaling, we may assume that |\Delta|\approx 1 . Although \Delta is not unique, we will choose a specific one and refer it as the ellipse associated with A . The family of balls \{\Delta_{k}\}_{k\in\mathbb{Z}} associated with A around the origin is defined by \Delta_{k}\coloneqq A^{k}(\Delta),\text{ where }k\in\mathbb{Z}. By Proposition 2.2 , we have \Delta_{k}\subset\Delta_{k+1} , and |\Delta_{k}|=b^{k}|\Delta_{0}|\approx b^{k} . Although all quasi-norms \rho_{A} are equivalent (see [ 2 ] , Chapter 1, Lemma 2.4), we will use the following specific one: From the nested property of the family \{\Delta_{k}\}_{k\in\mathbb{Z}} , we know that for every x\in\mathbb{R}^{d}\setminus\{0\} , there exists a unique k\in\mathbb{Z} so that x\in\Delta_{k+1}\setminus\Delta_{k} . We define \rho_{A}(x)\coloneqq\begin{cases}b^{k}&\quad\text{if }x\in\Delta_{k+1}\setminus\Delta_{k},\\ 0&\quad\text{otherwise}.\end{cases} Let A^{\ast} be the adjoint of A , since A^{\ast} and A have the same spectrum, the matrix A^{\ast} is also a dilation matrix. We can define the quasi-norm \rho_{A^{\ast}} , the nested balls \{\Delta^{\ast}_{k}\}_{k\in\mathbb{Z}} , and so on. 2.2. Anisotropic H^{1} , and the atomic decomposition Let \varphi\in\mathcal{S}(\mathbb{R}^{d}) with \int\varphi\neq 0 . The anisotropic dilation of \varphi is defined by \varphi_{k}(\cdot)\coloneqq b^{k}\varphi(A^{k}\cdot) . For f\in L^{1}(\mathbb{R}^{d}) , the smooth maximal function of f with respect to \varphi is defined by \mathcal{M}_{\varphi}(f)(x)\coloneqq\sup_{k\in\mathbb{Z}}|f\ast\varphi_{k}|(x). The anisotropic H^{1} space associated with the dilation A is defined by H^{1}_{A}(\mathbb{R}^{d})\coloneqq\Big\{f\in L^{1}(\mathbb{R}^{d}):\mathcal{M}_{\varphi}(f)\in L^{1}(\mathbb{R}^{d})\Big\} Similar to the isotropic case, the space H^{1}_{A} can be characterized by the atomic decomposition. We say that an L^{1} function a is an H^{1}_{A} - atom , if (1) There exist x_{0}\in\mathbb{R}^{d} and k\in\mathbb{Z} , so that \mathrm{supp}(a)\subset x_{0}+\Delta_{k} . (2) The function a is bounded by |\Delta_{k}|^{-1} almost everywhere: \|a\|_{\infty}\leq|\Delta_{k}|^{-1} . (3) The function a has vanishing moment: \int a=0 . The atomic decomposition of H^{1}_{A} states that f\in H^{1}_{A} if and only if there exists atoms \{a_{i}\}_{i=1}^{\infty} and a sequence (\lambda_{i})_{i=1}^{\infty}\in\ell^{1} , so that f=\sum_{i=1}^{\infty}\lambda_{i}a_{i} . Moreover, \displaystyle\|f\|_{H^{1}_{A}}\approx\inf\Big\{\sum_{i=1}^{\infty}|\lambda_{i}|:f=\sum_{i=1}^{\infty}\lambda_{i}a_{i}\Big\}, where the infimum is taken over all possible atomic decomposition of f . For this characterization, see [ 3 ] , Chapter 1, Theorem 6.5. 2.3. Duality, and anisotropic \mathrm{BMO} We denote \mathcal{B} the family of translated balls associated to A , \mathcal{B}\coloneqq\{x+\Delta_{k}:x\in\mathbb{R}^{d},\text{ and }k\in\mathbb{Z}\}. Assume that 1\leq q<\infty . A locally integrable function f is said to be in the anisotropic bounded mean oscillation space \mathrm{BMO}_{A,q} , if the following inequality holds: \displaystyle\sup_{B\in\mathcal{B}}\inf_{c_{B}\in\mathbb{C}}\bigg(\frac{1}{|B|}\int_{B}|f(x)-c_{B}|^{q}\,\mathrm{d}x\bigg)^{\frac{1}{q}}<\infty. This quantity also defines the \mathrm{BMO}_{A,q} norm. It is well-known that \mathrm{BMO}_{A,q} norms are equivalent for 1\leq q<\infty , so we will use the abbreviation \mathrm{BMO}_{A} and omit q . Also, as in the isotropic case, the following duality relation holds: (H^{1}_{A})^{\ast}=\mathrm{BMO}_{A}. For basic properties of the anisotropic \mathrm{BMO} , see [ 2 ] , Chapter 1, Section 8. 2.4. Use of notations and abbreviations For the later part of this article, we will drop the matrix A and A^{\ast} in the notations, which does not create ambiguity. For example, we use \rho,\rho_{\ast} to denote \rho_{A},\rho_{A^{\ast}} respectively. Unless otherwise specified, the spaces H^{1} and \mathrm{BMO} are always interpreted as H^{1}_{A} and \mathrm{BMO}_{A} . 3. Main theorem and its proof To state the main theorem, we need to introduce few more notations. For \Delta\subset\mathbb{R}^{d} the ellipse associated with A , we define R=R(\Delta) be the smallest rectangle contains \Delta . Here, for rectangles, we mean that the sides of R are parallel to the coordinate axes. By simple convex geometry, we know that |\Delta|\approx|R| , and there exists M=M_{\Delta}\in\mathbb{N} , so that A^{-M}(R)\subset\Delta\subset R. For \Delta^{\ast} , similar conclusion holds. We denote the smallest rectangle containing \Delta^{\ast} by R^{\ast} , and assume that (A^{\ast})^{-N}(R^{\ast})\subset\Delta^{\ast}\subset R^{\ast} , where N\in\mathbb{N} depends on the ellipse \Delta^{\ast} , and the matrix A^{\ast} . Suppose that R=I_{1}\times I_{2}\cdots\times I_{d} . For \alpha\in\mathbb{Z}^{d} , We denote R(\alpha) be the translation \displaystyle R(\alpha)\coloneqq\Big\{(x_{i})_{i=1}^{d}:|I_{i}|\alpha_{i}-\frac{|I_{i}|}{2}\leq x_{i}<|I_{i}|\alpha_{i}+\frac{|I_{i}|}{2}\Big\}, and \Delta(\alpha)\subset R(\alpha) be the corresponding translation of \Delta . Theorem 3.1 . Let \mu be a positive Borel measure