Asymmetric norms and optimal distance points in linear spaces (Q930745): Difference between revisions

Let \(X\) be a real vector space. A mapping \(q : X \to [0, \infty)\) is called an asymmetric norm and \((X,q)\) an asymmetric normed space if it satisfies, for all \(x,y \in X\) and \(a \geq 0\), the conditions \(q(a x) = a q(x),\; q(x+y) \leq q(x) + q(y)\), and \(q(x) = q(-x) = 0\) only for \(x = 0\). If \(q\) is an asymmetric norm, then \(q^s\) defined by \(q^s(x) = \max\{q(x), q(-x)\}\) is a norm. For a point \(x \in X\) and a set \(S \subseteq X\), denote by \(C_x(S)\) the set of all best approximations to \(x\) in \(C\), i.e., \(C_x(S) = \{y \in S : q(y-x) = \inf_{z \in S} q(z-x)\}\). A point \(y \in C_x(S)\) is called optimal distance point if, moreover, \(q^s(y-x) = \inf \{q^s(z-x) : z \in C_x(S)\}\). The authors study optimal distance points. In particular, they develop a technique to find optimal distance points using the notion of \(q\)-coverings.