The dual of the space of holomorphic functions on locally closed convex sets (Q2574027)

The article is a sequel to \textit{J.\ Bonet, R.\ Meise} and \textit{S.~N.\ Melikhov} [Bull.\ Belg.\ Math.\ Soc.\ -- Simon Stevin 10, No.~4, 491--503 (2003, Zbl 1086.46003)], and we refer to the review of this article for definitions, notation and background. The problem investigated in the previous as well as the present paper is to determine when for the weighted Hausdorff (LF)-space \({\mathcal V}H(\mathbb{C}^N)\) of entire functions, which arises via Laplace transform of the dual \(H(Q)'_b\) of the space \(H(Q)\) of all holomorphic germs on a locally closed convex set \(Q \subset \mathbb{C}^N\), projective description holds; i.e., when this space equals its projective hull \(H\overline{V}(\mathbb{C}^N)\) algebraically and/or topologically. If \(Q\) is bounded, in the previous article the authors derived necessary as well as sufficient conditions for algebraic and topological projective description. Here they treat the case when \(Q\) is not necessarily bounded. Among other things, they show: (1) If \(Q\) is, in addition, strictly convex at the relative boundary of \(\omega\); i.e., if the intersection of \(Q\) with each supporting complex hyperplane to \(\overline{Q}\) is compact, then projective description holds algebraically and topologically. (2) If \(Q\) has a neighborhood basis of domains of holomorphy and \({\mathcal V}H(\mathbb{C}^N)\) is a topological subspace of \(H\overline{V}(\mathbb{C}^N)\), then \(Q\) must be strictly convex at the relative boundary of \(\omega\). The problem which locally closed convex sets \(Q\) have a neighborhood basis of domains of holomorphy remains open, as does the problem of a characterization of the algebraic projective description for general locally closed convex sets \(Q\). The authors plan to return to the last problem in the future. The article also treats the corresponding projective description problem for spaces of continuous functions. Here, for any locally closed convex set \(Q\), \({\mathcal V}C(\mathbb{C}^N)\) is always a topological subspace of \(C\overline{V}(\mathbb{C}^N)\), and algebraic equality of the two spaces holds if and only if \(Q\) is strictly convex at the relative boundary of \(\omega\).