Index of symmetry and topological classification of asymmetric normed spaces (Q2219104)

An asymmetric norm on a real vector space \(X\) is a positively homogeneous subadditive functional \(\|\,{\cdot}\,|:X\to\mathbb{R}_+\) such that \(\|x|=\|{-}x|=0\) implies \(x=0\) (the possibility \(\|x|\ne \|{-}x|\) is not excluded). It generates in the usual way a \(T_0\) topology on \(X\), which is \(T_1\) iff \(\|x|>0\) whenever \(x\ne 0\). The functional \(\|x\|_s=\max\{\|x|,\|{-}x|\}\) is a norm on \(X\) and the corresponding normed space is denoted by \(X_s\) (see [\textit{S.~Cobzaş}, Functional analysis in asymmetric normed spaces. Basel: Birkhäuser (2013; Zbl 1266.46001)]). The authors introduce the index of symmetry of an asymmetric normed space \((X,\|\,{\cdot}\,|)\) by \(\, c(X):=\inf\{\|{-}x|: x\in X,\, \|x|=1\}\in [0,1]\). It turns out that it plays a crucial role in the study of the properties of the asymmetric normed space \(X\). For instance, if \(c(X)>0\), then \(X\) is \(T_1\) and \(c(X)\,\|x\|_s\le \|x|\le \|x\|_s\), which shows that the topology generated by the asymmetric norm is actually a norm topology -- the asymmetric normed space \((X,\|\,{\cdot}\,|)\) is isomorphic to \((X,\|\,{\cdot}\,\|_s)\). If \(X\) is finite-dimensional, then \(X\) is \(T_1\) iff \(c(X)>0\). If \((Y,\|\,{\cdot}\,|)\) is another asymmetric normed space, then the set \(L_c(X,Y)\) of all continuous linear operators from \(X\) to \(Y\) is a cone in the space \(L(X_s,Y_s)\) of all linear operators between the associated normed spaces \(X_s,Y_s\). On \(\mathbb{R}\) one considers the asymmetric norm \(\|t|_{\mathbb{R}}=\max\{0,t\}\) and one denotes by \(X^\flat\) the space \(L_c(X,\mathbb{R})\) (the asymmetric dual of \(X\)) and by \(X^*\) the dual of the associated normed space \(X_s\). The main concern of the paper is to study when \(L_c(X,Y)\) is not a vector subspace of \(L(X_s,Y_s)\). Among other results, it is shown that if \(L_c(X,Y)\) is not a vector space, then \(c(X)=c(Y)=0\). If, furthermore, \(Y\) is finite-dimensional, then the converse also holds (Corollary~1). Another problem studied in the paper is that of the \(w^*\)-density of \(X^\flat\) in \(X^*\). In general, span\((X^\flat)\) is \(w^*\)-dense in \(X^*\), and \(X^\flat\) is \(w^*\)-dense in \(X^*\) iff \(X\) is \(T_1\) (Theorem~2).