Existence of best approximation elements in \(C(Q, X)\) (Q1916612): Difference between revisions

Let \(X\) be a Banach space and let \(E\subset X\) be a linear subspace. For \(x\in X\) denote by \(P_E(x)\) the set of all best approximants for \(x\) in \(E\). It is said that \(E\) is a proximinal subspace of \(X\) if for every \(x\in X\) the set \(P_E(x)\) is nonempty. The aim of this paper is to characterize finite-codimensional proximinal subspaces of \(C(Q,X)\) (Theorems 2, 3, 4), where \(C(Q,X)\) denotes the space of all continuous functions defined on a compact set \(Q\) with values in a Banach space \(X\). In Theorems 2 and 3 it is assumed that \(X\) has a (wLD) property, i.e. if \(x,x_n\in S_X\), \(f\in S_{X^*}\), \(f(x)=1\), \(f(x_n)\to 1\) then \(x_n\) tends weakly to \(X\), where \(S_X\) (resp. \(S_{X^*}\)) is a unit sphere in \(X\) (resp. \(X^*\)). In Theorem 4, \(X\) is a finite-dimensional strictly convex space. Also, a characterization of finite-codimensional proximinal subspaces of an arbitrary Banach space \(B\) is presented (Theorem 1). The above results can be treated as a generalization of the results from the paper of A. L. Garkavi in which the case \(X=\mathbb{R}\) was considered (see also references included in the paper).