The Hadamard product as a universality preserving operator (Q977100): Difference between revisions

Let \(S\subset \mathbb{C}_{\infty}\) be a non-empty open set. We denote by \(\mathcal{G}(S)\) the set of all non-empty, open, simply connected subsets of \(\mathbb{C}\setminus S,\) by \(S^C\) its complement with respect to \(\mathbb{C} _{\infty}\), and by \(\mathcal{H}(S)\) the space of all holomorphic functions in \(S\) that vanish at infinity if \(\infty \in S.\) We consider an open, simply connected set \(D\subset \mathbb{C}\) with \(0 \in D\), and a set \(G\in \mathcal{G}(D).\) \textbf{Definition.} A function \(\varphi \in \mathcal{H}(D)\) is said to be universal with respect to \(G\) functions \((\varphi \in \mathcal{U}(D,G))\) if, for every \(f \in \mathcal{H}(G)\), there exists a sequence \((n_k)\) of non-negative integers so that \(s_{n_k}(\varphi) \to f\) locally uniformly on \(G\), where \(s_n(\varphi)\) is the \(n\)-th partial sum of the power series expansion of \(\varphi\) at \(0.\) Let \( \Omega\subset \mathbb{C}_ {\infty}\) be non-empty open set such that the set \(\Omega^C\cdot G^C\) does not contain terms of the form ``\(0\cdot \infty\)'' or ``\(\infty \cdot 0\)'', and such that \(\Omega^C\cdot G^C \neq \mathbb{C}_{\infty}\). Consider the set \(\Omega \ast G :=(\Omega^C\cdot G^C)^C\). Let \(\psi\in \mathcal{H}(\Omega)\), \(g\in \mathcal{H}(G)\). In their previous works, the authors have shown that one can find an appropriate integration curve \(\gamma \) in such a way that \[ \big(\psi \ast _{\Omega,G} g\big)(z) := \frac{1}{2\pi i} \int_{\gamma} \frac{\psi(\zeta)}{\zeta}g\left( \frac {z}{\zeta}\right) d\zeta \] defines a function in \(\mathcal{H}(\Omega \ast G )\). This function is called the Hadamard product of \(\psi\) and \(g\), and \[ \Big(\psi \ast _{\Omega, G} (g |_G)\Big)(z)=\sum _{\nu =0}^{\infty} \psi _{\nu} g_{\nu} z^{\nu} \] for all entire functions \(g\) (here \(\psi _{\nu}, g_{\nu}\) are the coefficients of the power series expansions at \(0\) for \(\psi\) and \(g\), respectively). Consider the operator \(\psi \ast _{\Omega, G}\cdot: \mathcal{H} (G)\rightarrow \mathcal{H}(\Omega \ast G )\) defined by \(\psi \ast _{\Omega, G} (g):=\psi \ast _{\Omega, G}g\). The following theorem is the main result of this paper. \textbf{Theorem.} (i) If the operator \(\psi \ast _{\Omega,G}\cdot: \mathcal{H}(G) \rightarrow \mathcal{H}(\Omega \ast G)\) has dense range, then \(\psi \ast _{\Omega , D}\phi \in \mathcal{U}(\Omega \ast D, \Omega \ast G)\) for every \(\phi \in \mathcal{U}(D \ast G).\) (ii) If \(\psi \ast _{\Omega , D}g \in \mathcal{U}(\Omega \ast D, \Omega \ast G)\) for \(g \in \mathcal{H}(D),\) then the operator \(\psi \ast _{\Omega , G}\cdot: \mathcal{H}(G) \rightarrow \mathcal{H}(\Omega \ast G)\) has dense range.