A geometric interpretation of best simultaneous approximations in conditional complete lattice Banach spaces (Q2836034)

The authors give a geometric interpretation of best simultaneous approximation from convex sets in conditional complete lattice Banach spaces with a strong unit and with norm derived by this strong unit.NEWLINENEWLINELet \(X\) be a normed linear space, let \(W\) and \(S\) be nonempty subsets of \(X\) with \(S\) bounded, let \( B(S,r):= \{x \in X : \sup_{s \in S}||s-x|| \leq r\}\), \( d(S,W):=\inf_{w \in W} \sup_{s \in S}|| s-w ||\) and \(P_W(S):=\{w \in W: \sup_{s \in S}||s - w || = d(S,W)\}\).NEWLINENEWLINEThe main result of this paper is the followingNEWLINENEWLINETheorem 3.2. Let \(W\) be a closed convex subset of \(X\), \(S\) a bounded set in \(X\), \(w_0 \in W\), and \(r:= \sup_{s \in S}||s-w_0||\). If \(||\sup S - w_0|| > ||\inf S-w_0||\) (resp., \(||\sup S - w_0|| < ||\inf S-w_0||\)), then the following statements are equivalent: (i)\ \(w_0 \in P_W(S)\); (ii)\ There exists an \(f \in X^{\star}\) with \(||f||=1\) such that the hyperplane \(H:= \{x \in X: f(x)=f(\sup S)-r\}\) (resp., \(H:= \{x \in X: f(x)=f(\inf S)-r\}\)) passes through \(w_0\) and separates \(W\) from \(B(S,r)\). Examples are provided.NEWLINENEWLINEThe authors state that this is a corrected version of Theorem 3.6 in [\textit{H. Mohebi} and \textit{E. Naraghirad}, Numer. Funct. Anal. Optim. 28, No. 11--12, 1327--1345 (2007; Zbl 1131.41006)].