Majorization in de Branges spaces. I: Representability of subspaces (Q961495)

\textit{L. de Branges} [``Some Hilbert spaces of entire functions'', Proc. Am. Math. Soc. 10, 840--846 (1959; Zbl 0092.07001)] initiated the study of Hilbert spaces of entire functions satisfying some additional axioms. These spaces can be viewed as a generalization of the classical Paley-Wiener spaces, which consist of all entire funcitons of exponential type at most \(a\) whose restriction to the real line is square integrable. The paper under review belongs to a series in which the authors investigate the aspect of majorization in de Branges spaces. Let \(\mathcal{H}\) be a de Branges space and \(m: D \to [0,\infty)\), where \(D \subset \mathbb{C}^+ \cup \mathbb{R}\). A function \(F \in \mathcal{H}\) is said to belong to \(R_m(\mathcal{H}) \) if there exists a constant \(C > 0\) such that \(\max\{|F(z)|, |F^{\sharp}(z)|\} \leq Cm(z)\) for any \(z \in D\), where \(F^{\sharp}(z) = \overline{F(\overline{z})}\). Then \(\mathcal{R}_m(\mathcal{H}) : = \text{clos}_{\mathcal{H}}(R_m(\mathcal{H}))\). Two questions motivate the authors in their investigation, namely, (a) which de Branges subspaces \(\mathcal{L}\) of a given de Branges space \(\mathcal{H}\) can be realized as \({\mathcal{L}}= {\mathcal{R}}_m ({\mathcal{H}})\) with some majorant \(m\)? (b) If \(\mathcal{L}\) is of the form \({\mathcal{R}}_m ({\mathcal{H}})\) with some \(m\), how big and how small can \(m\) be chosen so that still \({\mathcal{L}}= {\mathcal{R}}_m ({\mathcal{H}})\)? Up to now, only majorization along \(\mathbb{R}\) has been considered and the first question has been completely answered by the authors in Part II of this article series [Collectanea Mat. 62, No.~1, 27--55 (2011; Zbl 1262.46021)]. (Part III has appeared in [St. Petersbg. Math. J. 21, No.~6, 843--875 (2010); translation from Algebra Anal. 21, No.~6, 3--46 (2009; Zbl 1213.46023)].) The paper under review is focused to give some answers to these questions. As domain of majorization, the authors consider rays contained in the closed upper half-plane, lines parallel to the real axis contained in the closed upper half-plane, or a combination of such types of sets. After a preliminaries section, the main result in Section 3 is Theorem 3.1 which gives precise conditions on \(m\) for \({\mathcal{R}}_m({\mathcal{H}})\) to become a de Branges subspace of \(\mathcal{H}\). This theorem justifies the definition of admissible majorant for \(\mathcal{H}\) of which some examples are given. Since point evaluations are continuous in a de Branges space \(\mathcal{H}\), it is a reproducing kernel Hilbert space. For \(w \in \mathbb C\), let \(K(w,\cdot)\) be the kernel corresponding to \(w\) and let \(\nabla_{\mathcal{H}}(w): = \|K(w,\cdot)\|\). If \(\partial : \mathbb C \to \mathbb{N}_0\), set \(\text{Sub}_{\partial}{\mathcal{H}} = \{ \mathcal{L} \in \text{Sub}{\mathcal{H}} : \partial_{\mathcal{H}} = \partial \}\) and \(\text{Sub}^* \mathcal{H} = \text{Sub}_{\partial_{\mathcal{H}}} \mathcal{H}\). To a majorant \(m\) is associated its zero divisor \(\partial_m: \mathbb C \to \mathbb N \cup \{\infty\}\), where, for \(w \in \mathbb C\), \(\partial_m(w)\) is the infimum of all numbers \(n \in \mathbb N_0\) such that there exists a neighbourhood \(U\) of \(w\) with the property \[ \inf_{z \in U \cap D, z \neq w } \frac{m(z)}{|z - w|^n} > 0. \] Section 4 is dedicated to subspaces which are generated by majorization along rays not parallel to the real axis. In this respect, it is shown that if \(D = i[h, \infty)\), where \(h > 0\), and if \(\mathcal{L} \in \text{Sub}^{*} {\mathcal{H}}\), then \(\mathcal{L} = {\mathcal{R}}_{\nabla_{\mathcal{L}|D}}(\mathcal{H})\) (Theorem 4.1), and if \(D: = e^{i\pi \beta}[h, \infty)\), where \(h > 0\) and \(\beta \in (0, 1/2)\), each element of the de Branges space \(\mathcal{H}\) is of zero type with respect to the order \(\rho = (2 - 2\beta)^{-1}\) and \(\mathcal{L} \in \text{Sub}^{*}{\mathcal{H}}\), then \(\mathcal{L} = {\mathcal{R}}_{\nabla_{\mathcal{L}|D}}(\mathcal{H})\) (Theorem 4.2). In the last section, other results on majorization on sets which are close to the real axis are established. In short, a very well written paper which contains interesting results, some of them being sharp.