On various modes of scalar convergence in \(L_0({\mathfrak X})\) (Q5945952): Difference between revisions

From MaRDI portal
Normalize DOI.
Normalize DOI.
 
Property / DOI
 
Property / DOI: 10.1006/JMAA.2000.7459 / rank
Normal rank
 
Property / DOI
 
Property / DOI: 10.1006/JMAA.2000.7459 / rank
 
Normal rank

Latest revision as of 12:12, 9 December 2024

scientific article; zbMATH DE number 1657963
Language Label Description Also known as
English
On various modes of scalar convergence in \(L_0({\mathfrak X})\)
scientific article; zbMATH DE number 1657963

    Statements

    On various modes of scalar convergence in \(L_0({\mathfrak X})\) (English)
    0 references
    0 references
    0 references
    2001
    0 references
    Let \(X\) be a Banach space and \(L^p(X)\), \(1\leq p\leq\infty\), the spaces of Bochner integrable functions over [0,1] consisting of those \(X\)-valued strongly measurable functions with \(\| f(\cdot)\| _X\in L^p([0,1])\). The paper under review considers the problem of finding conditions, mainly on the Banach space \(X\), so that given any norm bounded sequence in \(L^p(X)\) satisfying some convergence properties we can extract a subsequence with other convergence properties. A sequence \((f_n)\) of \(X\)-valued strongly measurable functions is said to converge scalarly a.e. (resp. scalarly in measure; resp. scalarly in \(L^p\); resp. weakly a.e.) to a function \(f\) if for any \(x^*\in X^*\) the sequence \((x^*f_n)\) converges a.e. (resp. in measure; resp. in \(L^p\); resp. pointwise on \(A\), where \(A\) has full measure and is independent of \(x^*\)) to \(x^*f\). In the case \(p=\infty\), the authors show that the fact that \(L^\infty(X)\)-bounded sequences which converge scalarly a.e., or scalarly in measure, have a weakly a.e. convergent subsequence is equivalent to \(X^*\) having the Radon-Nikodym property. If we consider \(L^\infty(X)\)-bounded sequences which converge scalarly in measure and look for extracting scalarly a.e. convergent subsequences, the situation is more complicate. This is possible for \(X=\ell^1\) (a result given in personal communication to the authors by W. B. Johnson), but it is not true for \(X=C([0,1])\). The answer to this question for \(X=L^1\) is not known. In the case \(1\leq p<\infty\), it is shown, by a construction based on Dvoretzky's almost spherical sections theorem, that the fact that \(L^p(X)\)-bounded sequences which converge scalarly a.e., or scalarly in measure, or scalarly in \(L^p\), have a weakly a.e. convergent subsequence is equivalent to \(X\) being finite dimensional. If we consider \(L^p(X)\)-bounded sequences which converge scalarly in measure and look for extracting scalarly a.e. convergent subsequences, the situation is again more complicated. Some results are presented. In particular, this can be done for \(X=\ell^q\), \(1<q<\infty\), iff \(1<q'\leq p\leq\infty\); and for \(X=L^q\), \(1<q<\infty\), iff \(\max(2,q')\leq p\leq\infty\); \(q'\) being the conjugate exponent of \(q\).
    0 references
    Bochner spaces
    0 references
    strongly measurable functions
    0 references
    weak convergence
    0 references

    Identifiers

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