Extensions of p-adic vector measures (Q735426)

The author is one of the specialists in the theory of \(p\)-adic vector measures. This is the part of \(p\)-adic Functional Analysis devoted to the study of measures defined on a separating algebra \({\mathcal R}\) of subsets of a nonempty set \(X\) and with values in a Hausdorff locally convex space \(E\) over a complete non-archimedean and nontrivially valued field. Let \({\mathcal M}({\mathcal R},E)\) be the set of all finitely additive measures \(m: {\mathcal R} \rightarrow E\) such that \(m({\mathcal R})\) is a bounded subset of \(E\). For each such \(m\) and each continuous seminorm \(p\) on \(E\), we define \( m_p: {\mathcal R} \rightarrow \mathbb{R}, \, \, m_p(A) = \sup \{ p(m(V)): V \in {\mathcal R}, \, V \subset A \}.\) The main purpose of this paper is to give some results related to extensions of the following classes of measures. (a) \({\mathcal M}_{\sigma}({\mathcal R},E)\): the set of all \(m \in {\mathcal M}({\mathcal R},E)\) such that \(m(V_n) \to 0\) for each sequence \((V_n)\) in \({\mathcal R}\) with \(V_n \downarrow \varnothing\). (b) \({\mathcal M}_{\tau}({\mathcal R},E)\): the set of all \(m \in {\mathcal M}({\mathcal R},E)\) such that \(m(V_{\delta}) \to 0\) for each net \((V_{\delta})\) in \({\mathcal R}\) with \(V_ {\delta} \downarrow \varnothing\). By using some results of the previous papers of the author and some other ones obtained in the present paper, it is proved here that, for complete spaces \(E\), the following extension properties for the above classes of measures hold. (A) Let \(m \in {\mathcal M}_{\sigma}({\mathcal R},E)\) be strongly additive (i.e., for every sequence \((A_n)\) of pairwise disjoint members of \({\mathcal R}\), we have that \(m(A_n) \to 0\)). If \({\mathcal R}^{\sigma}\) is the \(\sigma\)-algebra generated by \({\mathcal R}\), then there exists a unique extension \(m^{\sigma} \in {\mathcal M}_{\sigma}({\mathcal R}^{\sigma},E)\) of \(m\). Moreover, for each continuous seminorm \(p\) on \(E\) and for each \(x \in X\), \( \inf \{ m_p(V): x \in V, \, V \in {\mathcal R} \} = \inf \{ m^{\sigma}_{p}(V): x \in V, \, V \in {\mathcal R} \}.\) (B) Let \(m \in {\mathcal M}_{\tau}({\mathcal R},E)\) be strongly additive. If \({\mathcal R}^{bo}\) is the \(\sigma\)-algebra of all \(\tau_{{\mathcal R}}\)-Borel sets in \(X\), then there exists a unique extension \(m^{\tau} \in {\mathcal M}_{\tau}({\mathcal R}^{bo},E)\) of \(m\). Moreover, \(m^{\tau} | {\mathcal R}^{\sigma} = m^{\sigma}\) (\(\tau_{{\mathcal R}}\) is the zero-dimensional Hausdorff topology on \(X\) having \({\mathcal R}\) as a topological base).