Geometric and probabilistic estimates for entropy and approximation numbers of operators (Q1104530)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Geometric and probabilistic estimates for entropy and approximation numbers of operators
scientific article

    Statements

    Geometric and probabilistic estimates for entropy and approximation numbers of operators (English)
    0 references
    1987
    0 references
    If X and Y are Banach spaces and \(T\in L(X,Y)\), the space of bounded linear mappings of X into Y, then \(e_ n(T)\) and \(a_ n(T)\), \(n\in {\mathbb{N}}\), are the sequences of entropy and approximation numbers of T. The principal results of the paper are the following. It is proved that if X and \(Y^*\) are of type 2 then there exist constants c, d, depending only upon X and Y, such that, for any \(T\in L(X,Y)\), \[ d^{-1} e_{[nc]}(T)\leq e(T^*)\leq de_{[n/c]}(T). \] This result generalizes one of \textit{B. Carl} [Friedrich Schiller Univ. Jena 30 and 31 (1983; Zbl 0585.47015 and Zbl 0585.47016)]. An asymptotic formula is obtained for the entropy numbers of diagonal operators on a real Banach space with a 1- unconditional basis. By a combination of probabilistic averaging, which is used to obtain bounds for approximation numbers, and a method of \textit{E. D. Gluskin} [Mat. Sb., Nov. Ser. 120(162), 180-189 (1983; Zbl 0528.46015)] estimates, asymptotic in k and n, are obtained for the approximation numbers \(a_ k(Id: c^ n_ p\to c^ n_{p'})\), \((1<p<2\), \(p'=p/p-1)\), where \(c^ n_ p\) is \(L(\ell^ n_ 2,\ell^ n_ 2)\) equipped with the norm \(\| T\| =(\sum_{i}a_ ui(T)^ p)^{1/p}\), and \(a_ k(Id: \ell^ n_ p\otimes_{\pi}\ell^ n_ p\to \ell^ n_ p\otimes_{\epsilon}\ell^ n_{p'})\) (1\(\leq p\geq 2)\).
    0 references
    space of bounded linear mappings
    0 references
    sequences of entropy and approximation numbers
    0 references
    asymptotic formula
    0 references
    diagonal operators
    0 references
    Banach space with a 1- unconditional basis
    0 references
    probabilistic averaging
    0 references
    0 references
    0 references
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references