Compact and precompact sets in asymmetric locally convex spaces (Q1019137)

It is well-known that the lack of the symmetry makes the theory of quasi-uniform spaces differ drastically from that of uniform spaces, mainly in what concerns completeness, compactness and total boundedness. Even the characterization of Hausdorff separation is more complicated in asymmetric spaces; this characterization was obtained for asymmetric normed spaces by \textit{L.\,M.\thinspace García-Raffi}, \textit{S.\,Romaguera} and \textit{E.\,A.\,Sánchez Pérez} in [Houston J.~Math.\ 29, No.\,3, 717--728 (2003; Zbl 1131.46300)] and extended for asymmetric locally convex spaces by \textit{S.\,Cobzaş} in [Int.\ J.\ Math.\ Math.\ Sci.\ 2005, No.\,16, 2585--2608 (2005; Zbl 1096.46002)], where basic definitions and properties of an asymmetric locally convex space \((X,P)\) were given; for instance, linear functional, dual space, weak topologies, extreme points, continuity of a linear mapping in terms of semi-Lipschitz conditions, separation of convex sets and the Krein-Milman theorem, etc. Here, \(X\) is a real vector space and \(P\) is a directed family of asymmetric seminorms on \(X\). In the present paper, the author continues the study of the basic properties of asymmetric locally convex spaces, obtaining that a \(T_{1}\) asymmetric locally convex space \((X,P)\) with dimension \(m\) is topologically isomorphic to the Euclidean space \(\mathbb{R}^{m}\). The aim of the present paper is to study precompactness and compactness within the framework of asymmetric locally convex spaces, extending results on compactness in asymmetric normed spaces obtained by \textit{L.\,M.\thinspace García-Raffi} in [Topology Appl.\ 153, No.\,5--6, 844--853 (2005; Zbl 1101.46017)] and by \textit{C.\,Alegre}, \textit{I.\,Ferrando}, \textit{L.\,M.\thinspace García-Raffi} and \textit{E.\,A.\thinspace Sánchez Pérez} in [Topology Appl.\ 155, No.\,6, 527--539 (2008; Zbl 1142.46004)]. Among others, the author proves here the following results: Let \((X,P)\) be an asymmetric locally convex space whose topology \(\tau (P)\) is \(T_{1}\). Then \(X\) is finite-dimensional if and only if there exists an outside \(P\)-precompact \(\tau (P)\)-neighborhood of \(0\). Recall that if \(p\) is an asymmetric seminorm on a vector space \(X\), a subset \(Y\) of \(X\) is outside \(p\)-precompact if, for every \(\varepsilon >0\), there exists a finite subset \(Z\) of \(X\) such that \(Y\subset \bigcup \left\{ B_{p}^{\prime }(z,\varepsilon ):z\in Z\right\} \), with \(B_{p}^{\prime }(z,\varepsilon ):=\left\{ y\in X:p(y-z)<\varepsilon \right\} \); therefore, a subset \(Y\) of an asymmetric locally convex space \((X,P)\) is outside precompact if for each asymmetric seminorm \(p\in P\) we have that \(Y\) is outside \(p\)-precompact. If \(Y\) is a precompact subset of an asymmetric locally convex space \((X,P)\), then the convex hull of \(Y\) is also precompact. A subset \(Y\) of an asymmetric locally convex space \((X,P)\) is \(\tau (P)\)-compact if and only if \(Y+\theta _{0}\) is \(\tau (P)\) is compact, where \(\theta _{0}=\bigcap \{\theta _{p}:p\in P\}\), with \(\theta _{p}=\{z\in X:p(z)=0\}\). This results are well-known in the locally convex classical theory, but in the asymmetric case the proofs are far from the classical ones. This paper contains a complete list of references, including the motivations related to the asymmetric cases that have been considered recently.