On Geometrical Properties of Metric Projections

Given a non-empty subset G of a metric space (X,d) and p ∈ X,anelement g0 ∈ Gis called a best approximation to p, if d(p,g0) = d(p,G) ≡ inf{d(p,g) : g ∈ G}. The map which takes each element p ∈ X to its set of best approximations in G is called a best approximation map or a metric projection. If G is a subspace of a linear metric space (X,d), we define the set G0 = {p ∈ X : d(p,0) = d(p,G)} of all those elements of X which have zero element as their best approximation in G. In this paper, we discuss some results concerning best approximation in metric spaces which are metrically convex, and the continuity of metric projection PG, where G is a Chebyshev subspace of a linear metric space, using the concept of G0 in linear metric spaces.

View article page

Download PDF