Convexity and best approximation in Banach spaces (Q2120260)

Let \(X\) be a real Banach space with dual \(X^*\) and let \(B_X\) and \(S_X\) be the closed unit ball and the unit sphere of \(X\), respectively. Let also \(J(x)=\{x^*\in S_{X^*}: x^*(x)=1\},\, x\in S_X\). For a nonempty subset \(C\) of \(X\) and \(x\in X\) let \(d(x,C)=\inf\{\|y-x\|: y\in C\}\) and \(P_C(x)=\{y\in C:\|y-x\|=d(x,C)\}\) be the distance of \(x\) to \(C\) and the metric projection on \(C\), respectively. The set \(C\) is called proximinal (Chebyshev) if \(P_C(x)\ne\emptyset\) (\(P_C(x) \) is a singleton) for every \(x\in C\). Obviously, every proximinal subset is closed. The set \(C\) is called approximatively compact (approximatively weakly compact), strongly Chebyshev (w-strongly Chebyshev) if every sequence \((y_n)\) in \(C\) such that \(\|y_n-x\|\to d(x,C)\) contains a subsequence convergent (weakly convergent) to some \(y\in C\), converges (converges weakly) to some \(y\in C\), respectively. Approximatively compact sets are approximatively weakly compact and these are proximinal. The space \(X\) is called strongly convex if for each sequence \((x_n)\) in \(S_X\) and \(x\in S_X,\, x_n\xrightarrow{\|\cdot\|}x\) whenever \(x^*(x_n)\to 1\) for some \(x^*\in J(x)\). It is called nearly strongly convex (very convex) if every such sequence contains a subsequence norm-convergent (weakly convergent) to \(x\). The paper is concerned with the connections between these two geometric properties of a Banach space and the proximinality properties of its closed convex subsets. For instance, for a nonempty convex subset \(C\) of a nearly strongly convex Banach space \(X\) the following are equivalent: \(C\) is proximinal, \(C\) is approximatively compact, \(C\) is approximatively weakly compact, \(C\) is strongly Chebyshev, \(C\) is w-strongly Chebyshev (Theorem 3.7). Also, a Banach space is very convex iff every convex proximinal subset of \(X\) is w-strongly Chebyshev iff every proximinal hyperplane in \(X\) is w-strongly Chebyshev (Theorem 4.7).