SAIN in a normed linear space (Q791070)

Let X be a normed linear space, M a dense linear subspace of X and \(\Gamma\) a finite subset of \(X^*\), the dual of X. The triple (X,M,\(\Gamma)\) has property SAIN if for each \(x\in X\) and each \(\epsilon>0\), there exists \(y\in M\) such that \(\| x-y\|<\epsilon, \gamma(x)=\gamma(y) \forall \gamma \in \Gamma\) and \(\| x\| =\| y\|\). Denoting by \(P_ L(x)\) the set of all best approximations of \(x\in X\) by means of L, the fact that L is semi-Chebyshev if \(P_ L(x)\) contains at most one element corresponding to every \(x\in X\) is exploited by the main theorem which states that if X is a real normed linear space and M is any dense convex subset of X, then (X,M,\(\Gamma)\) has property SAIN if and only if \(L^{\Theta}\subset M\), whenever L is a semi- Chebyshev subspace of X and \(L^{\Theta}\) is the set of all elements \(x\in X\) which are orthogonal to L (in the sense \(x\perp y\) if \(\| x+ty\| \geq \| x\|\), for every real number t). Another theorem concerning the utility of the concept of interpolatory elements of best approximation in this context marks the conclusion of this paper. The authors have since pursued this topic at great length and obtained quite a few results.