Lipschitz conditions, strong uniqueness, and almost Chebyshev subspaces of C(X) (Q797078)

Let C(X) denote the space of continuous, real-valued functions on the compact metric space X endowed with the uniform norm, and let M be a finite-dimensional subspace of C(X). For \(f\in C(X)\), let \(P_ M(f)\) denote the set of best uniform approximations to f from M. An element \(m^*\in M\) is said to be a unique best approximation to \(f\in C(X)\) if \(P_ M(f)=\{m^*\}\) and it is said to be a strongly unique best approximation to f if there is a constant \(r=r(f)>0\) such that \(\| f- m\| \geq \| f-m^*\| +r\| m-m^*\|\) for all \(m\in M\). The metric projection \(P_ M\) is said to be (Hausdorff) continuous at \(f\in C(X)\) if for every \(\epsilon>0\) there is a \(\delta>0\) such that \(h(P_ M(f),P_ M(g))<\epsilon\) for all \(g\in C(X)\) with \(\| f- g\|<\delta\), where h is the Hausdorff metric \([h(U,V)=\max \{\sup_{u\in U}\inf_{v\in V}\| u-u\|\), \(\sup_{v\in V}\inf_{u\in U}\| u-v\| \}\), U,V being non-empty closed subsets of C(X)]. \(P_ M\) is said to be point Lipschitz continuous at f if there is a constant \(\lambda =\lambda(f)>0\) such that \(h(P_ M(f)\), \(P_ M(g))\leq \lambda \| f-g\|\) for all \(g\in C(X).\) For a finite dimensional subspace M of C(X), X a compact metric space, it is well known that the (set-valued) metric projection \(P_ M\) is (Hausdorff) continuous at any \(f\in C(X)\) having a unique best approximation from M and is point Lipschitz continuous at any \(f\in C(X)\) having a strongly unique best approximation from M. The converses of these classical results are discussed in this paper. Main results: Theorem 1. Let \(f\in C(X)\). If f has a unique best approximation from M and \(P_ M\) is point Lipschitz continuous at f, then the best approximation to f is strongly unique. Theorem 2. Let M be an almost Chebyshev subspace of C(X) i.e. except for a set of first category in C(X) every function has a unique best approximation from M. (a) If \(P_ M\) is (Hausdorff) continuous at \(f\in C(X)\), then f has a unique best approximation from M. (b) If \(P_ M\) is point Lipschitz continuous at \(f\in C(X)\), then f has a strongly unique best approximation from M. Theorem 3. Let M be a finite dimensional subspace of C(X). Then the following are equivalent: (i) M is almost Chebyshev. (ii) If \(P_ M\) is continuous at f, then f has a unique best approximation from M. (iii) If \(P_ M\) is point Lipschitz continuous at f, then f has a strongly unique best approximation from M.