Nuclear embeddings in general vector-valued sequence spaces with an application to Sobolev embeddings of function spaces on quasi-bounded domains

From MaRDI portal
(Redirected from Publication:2074934)



Abstract: We study nuclear embeddings for spaces of Besov and Triebel-Lizorkin type defined on quasi-bounded domains OmegasubsetmathbbRd. The counterpart for such function spaces defined on bounded domains has been considered for a long time and the complete answer was obtained only recently. Compact embeddings for function spaces defined on quasi-bounded domains have been studied in detail already, also concerning their entropy and s-numbers. We now prove the first and complete nuclearity result in this context. The concept of nuclearity has been introduce by Grothendieck in 1955 already. Our second main contribution is the generalisation of the famous Tong result (1969) which characterises nuclear diagonal operators acting between sequence spaces of ellr type, 1leqrleqinfty. We can now extend this to the setting of general vector-valued sequence spaces of type with 1leqp,qleqinfty, MjinmathbbN0 and weight sequences with . In particular, we prove a criterion for the embedding to be nuclear.


The authors study nuclear operators on Besov and Triebel-Lizorkin spaces on quasi-bounded domains. Nuclear operators have recently been characterized for these spaces on bounded domains. The authors prove two results on nuclearity, one for embeddings of Besov and Triebel-Lizorkin spaces on unbounded domains and one for sequence spaces. If \(\Omega\) is a bounded Lipschitz domain in \(\mathbb R^d\), the identity map \(i:B^{s_1}_{p_1, q_1} (\Omega) \mapsto B^{s_2}_{p_2, q_2}(\Omega)\) is nuclear if and only if \[ s_1 - s_2 > d - d \max \left( \frac{1}{p_2} - \frac{1}{p_1} , 0 \right), \] where \(1 \leq p_i\), $q_i \leq \infty$, $s_i \in \mathbb R$, \(i = 1,2\). The authors wish to consider whether this might hold for unbounded domains, perhaps with nuclear replaced by compactness of the embedding. For unbounded domains, they show that if \(\Omega\) is unbounded and \(b(\Omega) = \infty\), then the embedding is nuclear if and only if \(p_1 = 1\), \(p_2 = \infty\), \(s_1 - s_2 > d\). If \(b(\Omega) < \infty\), the results look more like the bounded case. The embedding is nuclear if \[ s_1 - s_2 - d \left(\frac{1}{p_1} - \frac{1}{p_2} \right) > \frac{ b(\Omega) }{\mathbf{t}(p_1, p_2) } \] where \(1 \leq p_i, q_i \leq \infty\), \(s_i \in \mathbb R\), \(i = 1,2\), \(s_1 > s_2\), and \[\frac{1} {\mathbf{t}(p_1, p_2)} = 1 - \max \left(\frac{1}{p_1} - \frac{1}{p_2}, 0 \right) = 1 - \left(\frac{1}{p_1} - \frac{1}{p_2}\right)_+. \] \(b(\Omega)\) has a complicated technical definition, with \(b_j(\Omega)\) being the largest number of dyadic cubes of side length \(2^{-j}\) that are in \(\Omega\), and then \(b(\Omega) =\sup \{ t \in \mathbb R_+ : \limsup_{j \to \infty} b_j(\Omega) 2^{-jt} = \infty \}\). Note with the authors that for any non-empty open set \(\Omega \subset \mathbb R^d\), we have \(d \leq b(\Omega) \leq \infty\). If \(\Omega\) is unbounded and not quasi-bounded, then \(b(\Omega) = \infty\). But there are also quasi-bounded domains such that \(b(\Omega) = \infty\). Moreover, if the measure \(|\Omega |\) is finite, then \(b(\Omega) =d\) and one can check that when \(\Omega\) is bounded, the above result agrees with the first formula. The proof is based on a compactness result of \textit{A.~Tong} [Trans. Am. Math. Soc. 143, 235--247 (1969; Zbl 0186.45602)] for embeddings of \(\ell_p\), extended by \textit{H.-G. Leopold} [Georgian Math. J. 7, No.~4, 731--743 (2000; Zbl 0976.46025); in: Function spaces, differential operators and nonlinear analysis. Proceedings of the conference, FSDONA-99, Syöte, Finland, June 10--16, 1999. Prague: Mathematical Institute of the Academy of Sciences of the Czech Republic. 170--186 (2000; Zbl 0964.46011)] to the sequence space \(\ell_q ( \beta_j \ell_p^{M_j} )\). Leopold showed that if, with \(\beta_j >0\), \(M_j \in \mathbb N_0\), \(j \in \mathbb N_0\), \(M_j \geq 1\), \[ \ell_q ( \beta_j \ell_p^{M_j} )=\{ x = (x_{j,k})_{n \in \mathbb N_0, k = 1, \ldots, M_j}; x_{j,k} \in \mathbb C, \| x | \ell_q ( \beta_j \ell_p^{M_j} ) \| <\infty\} \] where \[ \| x | \ell_q ( \beta_j \ell_p^{M_j} )\| = \left( \sum_{j = 0}^{\infty} \beta_j^q \left( \sum_{k = 1}^{M_j} |x_{j,k}|^p \right)^{\frac{q}{p}} \right)^{\frac{1}{q}} , \] then the embedding of \(\ell_{p_1} ( \beta_j \ell_{p_2}^{M_j} )\) into \(\ell_{q_1} ( \ell_{q_2}^{M_j} )\) is compact if and only if \(( \beta_j^{-1} M_j^{\frac{1}{p^*}})_{j \in \mathbb N_0} \in \ell_{q^*}\), with \(\frac{1}{p^*} = \left(\frac{1}{p_2} - \frac{1}{p_1} \right)_+\), \(\frac{1}{q^*} = \left(\frac{1}{q_2} - \frac{1}{q_1} \right)_+\), modified if \(q^* = \infty\). Their result is based on their proof that the embedding of \(\ell_{p_1} ( \beta_j \ell_{p_2}^{M_j} )\) into \(\ell_{q_1} ( \ell_{q_2}^{M_j} )\) is nuclear if and only if \(( \beta_j^{-1} M_j^{\frac{1}{\mathbf{t}(p_1, p_2)}})_{j \in \mathbb N_0} \in \ell_{\mathbf{t}(q_1, q_2)}\), with \(\mathbf{t}(p,q)\) defined as above and modified if \(\mathbf{t}(q_1,q_2)= \infty\).



Cites work









This page was built for publication: Nuclear embeddings in general vector-valued sequence spaces with an application to Sobolev embeddings of function spaces on quasi-bounded domains

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2074934)