Simply interpolating and Carleson sequences for Hardy spaces in the polydisc
Nikolaos Chalmoukis, Alberto Dayan · 2024
Read Simply interpolating and Carleson sequences for Hardy spaces in the polydisc by Nikolaos Chalmoukis, Alberto Dayan (2024) 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-2412-09099.txt.
Introduction
Nikolaos Chalmoukis and Alberto Dayan study in 2024 sequences that can steer the values of analytic functions on the polydisc, a domain shaped like several discs multiplied together. Interpolation asks whether a function can be forced to take chosen values at chosen points. Two strengths of the property, simply and universally interpolating, are at issue.
For one disc the two properties coincide, by a classical theorem. In higher dimensions the picture splits by the exponent of the Hardy space. For some ranges the theorem survives. But for exponents between one and two, the authors show the two properties come apart. Their counterexample settles that range.
Catalog record
| CREATOR | Nikolaos Chalmoukis, Alberto Dayan |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 38 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-09099
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Simply interpolating and Carleson sequences for Hardy spaces in the polydisc Thanks: N. Chalmoukis is a member of the Gruppo Nazionale per l’Analisi Matematica, la Probabilità e le loro Applicazioni (GNAMPA) of the Istituto Nazionale di Alta Matematica (INdAM) Thanks: A. Dayan is partially supported by the Emmy Noether Program of the German Research Foundation (DFG Grant 466012782) Nikolaos Chalmoukis, Alberto Dayan Abstract . We study the relation between simply and universally interpolating sequences for the holomorphic Hardy spaces H p ( 𝔻 d ) H^{p}(\mathbb{D}^{d}) on the polydisc. In dimension d = 1 d=1 a sequence is simply interpolating if and only if it is universally interpolating, due to a classical theorem of Shapiro and Shields. In dimension d ≥ 2 d\geq 2 , Amar showed that Shapiro and Shields’ theorem holds for H p ( 𝔻 d ) H^{p}(\mathbb{D}^{d}) when p ≥ 4 p\geq 4 . In contrast, we show that if 1 ≤ p ≤ 2 1\leq p\leq 2 there exist simply interpolating sequences which are not universally interpolating. 2020 Mathematics Subject Classification Primary 32A35; Secondary 32E30 1. Introduction Let f be a holomorphic function in the d -dimensional polydisc \mathbb{D}^{d} . We say that f belongs to the Hardy space H^{p}(\mathbb{D}^{d}) , if \|f\|_{p}^{p}:=\sup_{0<r<1}\int_{\mathbb{T}^{d}}|f(rz)|^{p}dm(z)<+\infty, where \mathbb{T}^{d} is the d -dimensional torus and m is its probability Haar measure. We will also occasionally use the notation |\,\,\,| to denote m. The spaces of bounded holomorphic functions H^{\infty}(\mathbb{D}^{d}) is defined analogously. The present paper grew out of the attempt to better understand interpolating sequences for Hardy spaces in the higher dimensional polydisc. In dimension one, the work of Carleson [ 6 ] and Shapiro and Shields [ 18 ] provides a solid understanding of interpolating sequences. Let us spend a few words to recall the basic definitions and results in this simpler case. Consider \Lambda=(\lambda_{n})_{n} a sequence of points in the unit disc \mathbb{D} . Then (1-|\lambda_{n}|^{2})^{\frac{1}{p}}|f(\lambda_{n})|\leq\|f\|_{p},\,\,\,f\in H^{p}(\mathbb{D}), hence for 1\leq p<\infty the restriction operator at \Lambda on H^{p}(\mathbb{D}) is naturally defined as T^{p}_{\Lambda}(f)=\big(f(\lambda_{n})(1-|\lambda_{n}|^{2})^{\frac{1}{p}}\big),\,\,\,f\in H^{p}(\mathbb{D}). A sequence is called simply interpolating if T_{\Lambda}^{p}(H^{p}(\mathbb{D}))\supseteq\ell^{p} . Explicitly, if for every sequence of data (a_{n})\in\ell^{p} there exists an interpolating function f\in H^{p}(\mathbb{D}) , in the sense that (1.1) f(\lambda_{n})=a_{n}(1-|\lambda_{n}|^{2})^{-\frac{1}{p}},\,\,n=1,2,\dots It is worth mentioning that, since all Hardy spaces contain the polynomials, an interpolation problem like ( 1.1 ) always has a solution if the number of nodes is finite. If \Lambda is a simply interpolating, an application of the open mapping theorem yields that there exists a constant C , depending only on \Lambda and p , such that one can find a solution f for ( 1.1 ) such that \|f\|_{p}\leq C\|(a_{n})\|_{\ell^{p}} (see Lemma 2.1 ). A sequence is called universally interpolating if it is simply interpolating and T_{\Lambda}^{p} maps boundedly H^{p}(\mathbb{D}) into \ell^{p}. In other words a sequence is universally interpolating if for every \ell^{p} data the interpolation problem ( 1.1 ) has a solution and there exists M>0 such that (1.2) \sum_{n=1}^{\infty}|f(\lambda_{n})|^{p}(1-|\lambda_{n}|^{2})\leq M\|f\|^{p}_{p},\,\,\forall f\in H^{p}(\mathbb{D}). Sequences that satisfy ( 1.2 ), are usually called Carleson sequences . If \mu is a positive Borel measure in \mathbb{D} we say that it is an H^{p}(\mathbb{D}) - Carleson measure if there exists M>0 such that (1.3) \int_{\mathbb{D}}|f|^{p}d\mu\leq M\|f\|^{p}_{p},\,\,\,\forall f\in H^{p}(\mathbb{D}). Hence \Lambda is a Carleson sequence if and only if the atomic measure \mu_{\Lambda}:=\sum_{n}(1-|\lambda_{n}|^{2})~\delta_{\lambda_{n}} supported on \Lambda is a Carleson measure. It is a classical theorem of Carleson that the class of the homonymous measures is the same for every p\geq 1 and a measure is such if and only if there exists C>0 such that (1.4) \mu(S(I))\leq C|I|, for every arc I\subseteq\mathbb{T} and every region S(I) of the form S(I)=\{z\in\mathbb{D}\setminus\{0\}:1-|z|\leq|I|,z/|z|\in I\}. Furthermore, Shapiro and Shields showed that a sequence is universally interpolating for H^{p}(\mathbb{D}) if it is hyperbolically separated , that is \inf_{n\neq m}\Big|\frac{\lambda_{n}-\lambda_{m}}{1-\overline{\lambda_{m}}\lambda_{n}}\Big|>0 and it is a Carleson sequence. Even more surprisingly, they proved that simply interpolating sequences are the same for all p\geq 1 and that a simply interpolating sequence for H^{p}(\mathbb{D}) is automatically universally interpolating. In other words, if for some sequence \Lambda\subseteq\mathbb{D} we have T_{\Lambda}^{p}(H^{p}(\mathbb{D}))\supseteq\ell^{p} then T_{\Lambda}^{p}(H^{p}(\mathbb{D}))=\ell^{p}. In dimension d\geq 2 the analogous problems concerning H^{p}(\mathbb{D}^{d}) functions can be stated using the following restriction operator. Let \Lambda=(\lambda_{n})\subseteq\mathbb{D}^{d},\,\,\lambda_{n}=(\lambda_{n}^{1},\dots,\lambda_{n}^{d}) . Then define (1.5) T_{\Lambda}^{p}(f)=\Big(f(\lambda_{n})\prod_{i=1}^{d}(1-|\lambda_{n}^{i}|^{2})^{\frac{1}{p}}\Big),\qquad f\in H^{p}(\mathbb{D}^{d}). If p=\infty , T^{\infty}_{\Lambda} is the unweighted restriction operator on H^{\infty}(\mathbb{D}^{d}) . We say that a sequence \Lambda is; (SI) simply interpolating for H^{p}(\mathbb{D}^{d}) if T_{\Lambda}^{p}(H^{p}(\mathbb{D}^{d}))\supseteq\ell^{p} , (UI) universally interpolating for H^{p}(\mathbb{D}^{d}) if T_{\Lambda}^{p}(H^{p}(\mathbb{D}^{d}))=\ell^{p} , (CS) a Carleson sequence for H^{p}(\mathbb{D}^{d}) if T_{\Lambda}^{p}(H^{p}(\mathbb{D}^{d}))\subseteq\ell^{p}. Similarly, a positive Borel measure on \mathbb{D}^{d} is called Carleson if H^{p}(\mathbb{D}^{d})\subseteq L^{p}(\mathbb{D}^{d},\mu). Then \Lambda is a Carleson sequence if and only if the atomic measure (1.6) \mu_{\Lambda}:=\sum_{n}\left(\prod_{i=1}^{d}1-|\lambda_{n}^{i}|^{2}\right)~\delta_{\lambda_{n}} is a Carleson measure. As in the one dimensional case if \Lambda is a simply interpolating sequence there exists C>0 such that for every a\in\ell^{p} , there exists f\in H^{p}(\mathbb{D}^{d}) such that T_{\Lambda}^{p}(f)=a and \|f\|_{p}\leq C\|a\|_{\ell^{p}} . The infimum of such C is the simple interpolation constant of \Lambda . Similarly, given a Carleson sequence \Lambda we define its Carleson constant as the norm of T^{p}_{\Lambda}:H^{p}(\mathbb{D}^{d})\to\ell^{p} . The corresponding separation condition, also called weak separation becomes \inf_{n\neq m}\max_{1\leq i\leq d}\Big|\frac{\lambda^{i}_{n}-\lambda^{i}_{m}}{1-\overline{\lambda^{i}_{n}}\lambda^{i}_{m}}\Big|>0. A first hint for how two or more dimensions are different than one came again from Carleson [ 7 ] . He considered the Hardy space h^{p}(\mathbb{D}^{d}) of separately harmonic functions, which consists of functions u\in C^{\infty}(\mathbb{D}^{d}) such that \frac{\partial^{2}u}{\partial\overline{z_{i}}\partial z_{i}}=0,1\leq i\leq d\,\,\,\text{and}\,\,\,\,\|u\|^{p}_{p}=\sup_{0<r<1}\int_{\mathbb{T}^{d}}|u(rz)|^{p}dm(z)<+\infty. In d=1 , it is an immediate consequence of the M. Riesz Theorem that if p>1 , then H^{p}(\mathbb{D})\subseteq L^{p}(\mathbb{D},\mu) if and only if h^{p}(\mathbb{D})\subseteq L^{p}(\mathbb{D},\mu) . Consequently, a quite natural conjecture would be that h^{p}(\mathbb{D}^{d})\subseteq L^{p}(\mathbb{D}^{d},\mu) if and only if the d -dimensional version of ( 1.4 ), namely \mu(S(I_{1})\times\dots S(I_{d}))\leq C|I_{1}|\cdots|I_{d}|, holds for all arcs I_{1},\dots I_{d}\subseteq\mathbb{D} and some fixed constant C>0. Often this condition is referred to as the one box condition. It turns out that this is not the case. In fact, there exists a measure \mu satisfying the one box condition but h^{p}(\mathbb{D}^{2})\not\subseteq L^{p}(\mathbb{D}^{2},\mu) for any p>1 . M As an intermediate step towards our results on interpolation, via a quite elementary modification of Carleson’s example, we can show a little bit more; there exists a measure \mu in \mathbb{D}^{2} satisfying the one box condition but H^{p}(\mathbb{D}^{2})\not\subseteq L^{p}(\mathbb{D}^{2},\mu) for any p\geq 1 . The details are given in Section 5 . Nonetheless, a complete characterization of Carleson measures for the holomorphic Hardy spaces in higher dimensions has yet to emerge. In particular it is not clear whether the sufficient condition of Chang - Stein [ 11 ] is also necessary, or if the class of H^{p}(\mathbb{D}^{d}) Carleson measures is the same for every p\geq 1 for that matter. We remark that often in the literature the necessity of the Chang - Stein condition has been assumed to be true, but to the best of our knowledge a proof is not known. The Chang-Stein condition characterizes Carleson measures for all Hardy spaces of separately harmonic functions on the polydisc. By inclusions, any Carleson measure for a separately harmonic space is Carleson for its holomorphic counterpart, but the converse implication is not known in dimension d\geq 2 . More details are provided in