A characterization of best simultaneous approximations (Q5899755)

Let X be a compact Hausdorff space and Y a normed linear space with norm \(\| \cdot \|\). Let C(X,Y) denote the set of all continuous functions from X to Y. Suppose that functions \(F_ 1,...,F_{\ell}\) in C(X,Y) are given. The problem is to approximate these functions simultaneously by functions in S, an n-dimensional subspace of C(X,Y), in the sense of Chebyshev. That is, to find a function \(f\in S\) which minimizes \(\max_{1\leq j\leq \ell}\max_{x\in S}\| F_ j(x)- f(x)\|\) over the set S. If such a function \(f^*\) in S exists, they call it a best simultaneous approximation for \(F_ 1,...,F_{\ell}\). In this note a necessary and sufficient condition for a function to be a best simultaneous approximation for \(F_ 1,...,F_{\ell}\) is given.