scientific article; zbMATH DE number 3927580

From MaRDI portal
Publication:3701910





Let E be a locally K-convex space over a non-trivially valued field whose topology is given by a family \(\Gamma\) of non-Archimedean semi-norms. Let G be a subset of E, \(x\in E/G\) and \(g_ 0\in G\). Then \(g_ 0\) is called (i) an element of best approximation of x by means of elements of G if for every \(p\in \Gamma\), \(p(x-g_ 0)\leq p(x-g)\) for all \(g\in G\), (ii) an element of best coapproximation of x by means of elements of G if for every \(p\in \Gamma\), \(p(g_ 0-g)\leq p(x-g)\) for all \(g\in G\). An element \(x\in E\) is said to be orthogonal to \(y\in E\) if p(x)\(\leq p(x-\alpha y)\) for every scalar \(\alpha\) and for every \(p\in \Gamma\), x is said to be orthogonal to G if \(x\perp y\) for all \(y\in G\). In this paper, some characterizations of elements of best approximation and best coapproximation are given, and using the notion of orthogonality it is shown that there is no need to study best coapproximation in the above locally K-convex spaces separately as in these spaces the elements of best approximation and best coapproximation coincide. Main Theorem: Let G be a linear subspace of E and \(x\in E| G\). Then an element of best approximation to x in G is an element of best coapproximation and vice-versa.












This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3701910)