Some new bounds on the entropy numbers of diagonal operators (Q2291480)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 7160914
Language Label Description Also known as
default for all languages
No label defined
    English
    Some new bounds on the entropy numbers of diagonal operators
    scientific article; zbMATH DE number 7160914

      Statements

      Some new bounds on the entropy numbers of diagonal operators (English)
      0 references
      0 references
      31 January 2020
      0 references
      Let \(0<p,q\leq \infty\) and \(\sigma= (\sigma_k)_k\) be a non-negative and non-increasing sequence of real numbers. The diagonal operator associated to \(\sigma\), \(D_\sigma: \ell_p \to \ell_q\) is defined as \(D((x_n)_n):= (\sigma_n x_n)_n\). The entropy numbers associated to \(D_\sigma\) are defined by \[\varepsilon_n(D_\sigma):= \inf \left\{ \varepsilon>0: \text{there exist } y_1, \dots, y_n\in \ell_q \text{ with } D_\sigma(B_{\ell_p}) \subseteq \bigcup_{i=1}^n y_i + \varepsilon B_{\ell_q}\right\}.\] The main theorems claim that, for \(0<p<q \leq \infty\) with \(1/p=1/q+1/s\), and \(\sigma\) a sequence of positive numbers decreasing to zero, the entropy of \(D_\sigma\) satisfies the estimate \[\varepsilon_n(D_\sigma) \leq c \cdot \sup_{k \geq 1} k^{-1/s} \left( \frac{(\sigma_1+k^{1/s}\sigma_k) \dots (\sigma_k+k^{1/s}\sigma_k)}{n}\right)^{1/k}.\] Moreover, if there is \(b>1\) such that \(\sup_{k \leq n} \frac{\sigma_n b^n}{\sigma_k b^k} < \infty\) (where the supremum is over all \(n, k \in \mathbb{N}\), \(k\leq n\)), then the above bound is optimal. On the other hand, in the case where \(0<q<p \leq \infty\) with \(1/p=1/q+1/r\), with \(\sigma \in \ell_r\) a sequence of positive numbers decreasing to zero, the entropy of \(D_\sigma\) satisfies the estimate \[\varepsilon_n(D_\sigma) \leq c \cdot \sup_{k \geq 1} \left( \frac{(\tau_k+k^{1/r}\sigma_1) \dots (\tau_k+k^{1/r}\sigma_k)}{n}\right)^{1/k},\] where \(\tau_k:=(\sum_{n=k}^\infty \sigma_n^r)^{1/r}\). Moreover, under suitable assumptions (either (ALP) or (AMP)), the previous bound is optimal.
      0 references
      diagonal operators
      0 references
      entropy numbers
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references

      Identifiers