Convergence of best approximations in a smooth Banach space (Q797076)

From MaRDI portal
Revision as of 12:41, 14 June 2024 by ReferenceBot (talk | contribs) (‎Changed an Item)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
scientific article
Language Label Description Also known as
English
Convergence of best approximations in a smooth Banach space
scientific article

    Statements

    Convergence of best approximations in a smooth Banach space (English)
    0 references
    0 references
    1984
    0 references
    Let X be a reflexive and strictly convex Banach space. Then for any non- empty closed convex subsets \(C\subseteq X\) and \(x\in X\) there exists a unique best approximation \(p(x| C)\) of x by elements of C. Whenever the sequence \(\{x_ n\}\subseteq X\) which weakly converges to some \(x\in X\) and \(\| x_ n\| \to \| x\|\) as \(n\to \infty\) necessarily converges to x in norm, X is said to have Property (H). If X further has Property (H), then the mapping \(x\mapsto p(x| C)\) is continuous in norm. In this paper, the continuity of mapping \(C\mapsto p(x| C)\) is investigated with the non-parametralized method. Let \(\{C_ n\}\) be a sequence of non-empty closed convex subsets of X. The following is a main theorem. If X has Property (H), then for any \(x\in X\) the sequence \(\{p(x| C_ n)\}\) converges to \(p(x| \lim C_ n)\) in norm whenever lim \(C_ n\) exists and is not empty. Conversely, if X is finite dimensional or has Fréchet differentiable norm, then lim \(C_ n\) (the strong limit of \(\{C_ n\}\) in Mosco's sense) exists and is not empty whenever the sequence \(\{p(x| C_ n)\}\) converges in norm for every \(x\in X\). Using the limit of closed convex sets, a measurability condition of closed convex set valued functions can be defined. By the above result, in a certain Banach space, the condition is equivalent to some measurability conditions which are known in the theory of multi- valued functions.
    0 references
    unique best approximation
    0 references
    non-parametralized method
    0 references
    multi-valued functions
    0 references

    Identifiers