Strong proximinality of closed convex sets (Q2431557)

Let \(X\) be a Banach space with unit ball \(B_X\), unit sphere \(S_X\) and dual space \(X^*\). For \(f\in X^*\), with \(\|f\|=1\), set \(J_X(f)=\{ x\in S_X: f(x)=1\}\). The functional \(f\) is norm attaining if \(J_X(f)\neq \emptyset\). A closed set \(C\) is strongly proximinal if, for each \(x\in X\) and \(\varepsilon>0\), there exists \(t>0\) such that \(P_C(x,t)\subset P_C(x)+\varepsilon B_X\), where \(P_C(x)=\{ y\in C: \|x-y\|= d(x,C)\}\) and \(P_C(x,t)=\{ y\in C: \|x-y\|< d(x,C)+t\}\). The set \(C\) is approximatively compact if every minimizing sequence in \(C\) has a convergent subsequence. The norm \(\|\cdot\|\) is said to be strongly sub-differentiable (SSD), if for every \(x\in S_X\) the limit \[ \lim_{t\to 0^+}\frac{\|x+th\|-\|x\|}{t} \] exists uniformly for \(h\in S_X\). The main result of the paper is the following: Theorem. For a Banach space \(X\) the following assertions are equivalent: {\parindent=7,5mm \begin{itemize}\item[(i)] The norm of \(X^*\) is strongly sub-differentiable and, for every \(f\in S_{X^*}\), the set \(J_X(f)\) is compact. \item[(ii)] \(X\) is reflexive and the relative weak and the norm topologies coincide on the unit sphere \(S_X\) of \(X\). \item[(iii)] Every closed convex subset of \(X\) is approximatively compact. \item[(iv)] Every closed convex subset of \(X\) is strongly proximinal. \end{itemize}}