On a property of PLS-spaces inherited by their tensor products (Q961982)

From MaRDI portal
Revision as of 23:04, 7 July 2023 by Importer (talk | contribs) (‎Created a new Item)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
scientific article
Language Label Description Also known as
English
On a property of PLS-spaces inherited by their tensor products
scientific article

    Statements

    On a property of PLS-spaces inherited by their tensor products (English)
    0 references
    0 references
    1 April 2010
    0 references
    The aim of this paper is to investigate when the dual interpolation estimate is inherited by tensor products of PLS-spaces. The class of PLS-spaces is the smallest class of locally convex spaces containing all duals of Fréchet-Schwartz spaces and closed with respect to countable products and closed subspaces. This class also contains all Fréchet-Schwartz spaces and the most relevant spaces which appear in applications of linear functional analysis, such as spaces of real analytic or quasi-analytic functions, spaces of holomorphic or smooth functions, as well as spaces of (ultra)distributions. Every PLS-space \(X\) is of the form \(X=\, \text{proj}_{N\in {\mathbb N}}\, \text{ind }_{n\in {\mathbb N}}\,X_{N,n},\) where all the \(X_{N,n}\) are Banach spaces with unit ball \(B_{N,n},\) and the linking maps \(\iota^N_n:X_{N,n} \to X_{N,n+1} \) are compact. Hence \(X_N:=\text{ind }_{n\in {\mathbb N}}\,X_{N,n}\) are duals of Fréchet-Schwartz spaces. Denote by \(\iota_N:X \to X_N\) the canonical projection. The PLS-space \(X\) is said to have the dual interpolation estimate for small \(\theta\) if \[ \begin{multlined} \forall N \, \exists M\, \forall K\, \exists n \, \forall m \, \exists \theta_0 \in (0,1) \, \forall \,\theta \leq \theta_0\, \exists k,\,C >0 \, \forall x' \in X'_N \, : \\ \|x' \circ \iota ^M_N\|^*_{M,m} \leq C( \|x' \circ \iota ^K_N\|^*_{K,k})^{1-\theta}(\|x' \|^*_{N,n})^{\theta}. \end{multlined} \] If the inequality is satisfied for all \(\theta \geq \theta_0\) (resp., for all \(\theta \in (0,1)\)), then \(X\) has the dual interpolation estimate for big \(\theta \) (resp., for all \(\theta\)). The dual interpolation estimates where introduced by \textit{J.\,Bonet} and \textit{P.\,Domanski} in [Arch.\ Math.\ 89, 430--441 (2007; Zbl 1147.46024)] and they unify the four (\(\Omega\))-type invariants (P\(\Omega\)), (P\(\overline{\overline {\Omega}}\)), (P\(\underline{A}\)) and (P\(\underline{\underline{A}}\)), previously defined by the same authors. After an expository section which contains all definitions and properties needed in the paper, the author shows the main results in Section 3. In particular, he shows that the dual interpolation estimate for all \(\theta\) is inherited by the projective and injective tensor products of two Fréchet-Schwartz spaces as well as of their duals (Theorems 4 and 5). If one of the spaces is nuclear and the other is an arbitrary PLS-space, all interpolation estimates are inherited by their completed tensor product (Theorem 6).
    0 references
    Fréchet-Schwartz space
    0 references
    PLS spaces
    0 references
    (DN) and (\(\Omega\)) type conditions
    0 references
    dual interpolation estimate
    0 references
    tensor products
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references