Wavelet para-bases and sampling numbers in function spaces on domains

DOI10.1016/J.JCO.2006.08.002MaRDI QIDQ2465283zbMATH OpenOpenAlexFDO

Publication date 9 January 2008

Published in Journal of Complexity (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1016/j.jco.2006.08.002

function spaces polynomial reproducing formulas sampling numbers wavelets on domains

Rate of convergence, degree of approximation (41A25) Approximation by arbitrary nonlinear expressions; widths and entropy (41A46) Function spaces arising in harmonic analysis (42B35) Nontrigonometric harmonic analysis involving wavelets and other special systems (42C40) Sobolev spaces and other spaces of ``smooth functions, embedding theorems, trace theorems (46E35)

Review text

Let \(\Omega\) be an open set (domain) in \(\mathbb{R}^n\). There are different possibilities to consider function spaces of type \(B^s_{pq}\) or \(F^s_{pq}\) on \(\Omega\), where \(s\in\mathbb{R}\), \(0<p,q\leq\infty\) (with \(p<\infty\) in case of \(F^s_{pq}\)). In the present paper several approaches are discussed in detail, covering in particular those obtained by restriction from their \(\mathbb{R}^n\)-counterparts, by intrinsic definitions, by completion (and restriction) of compactly supported (in \(\Omega\)) smooth functions, as well as adapted to a refined Whitney decomposition in \(\Omega\). Main emphasis is put on very general domains and so-called \(E\)-thick domains (including, in particular, Lipschitz domains as well as many self-similar fractals). A second central topic are wavelet decompositions of such spaces, used to prove diversity and coincidences of the spaces under consideration, but also interesting on its own. The same applies to the local polynomial reproducing formulas dealt with in Section~4. Finally, in Section~5, sampling numbers of compact embeddings of the type \[ \widetilde{\mathrm{id}}_\Omega : \widetilde{B}^s_{pq}(\Omega) \hookrightarrow L_t(\Omega), \quad 0<t\leq\infty, \] are studied, where \(0<p,q\leq\infty\), \(s>\frac{n}{p}\), and \(\Omega\) stands for an arbitrary or \(E\)-thick domain in \(\mathbb{R}^n\) with finite measure. It turns out that asymptotically the sampling numbers \(g_k(\widetilde{\mathrm{id}}_\Omega)\) behave like \(k^{-s/n + \max(1/p-1/t,0)}\), \(k\in\mathbb{N}\), no matter whether linear or non-linear sampling numbers are considered.

Recommendations

Cites work

Cited in

(3)

This page was built for publication: Wavelet para-bases and sampling numbers in function spaces on domains

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2465283)