Splitting the Riesz basis condition for systems of dilated functions through Dirichlet series (Q2247708)

Each square-integrable function \(\varphi\) on the intervall \((0,1)\) (considered as a \(2\)-periodic odd function defined on \(\mathbb{R}\)) defines the system \((\varphi(n x))_{n \in \mathbb{N}}\) of dilations in the Hilbert space \(L_2(0,1)\). The article [\textit{H. Hedenmalm} et al., Duke Math. J. 86, No. 1, 1--37 (1997; Zbl 0887.46008)] characterizes those functions \(\varphi \) for which the associated system of dilations forms a Riesz basis in \(L_2(0,1)\). More precisely, following ideas of Beurling and Bohr, the main result of Hedenmalm, Linqvist and Seip [loc.\,cit.]\ reads as follows: Consider some function \(\varphi = \sum_n a_n \sqrt{2} \sin(\pi n x)\) (recall that \((\sqrt{2} \sin(\pi n x)_n\) forms an orthonormal basis of \(L_2(0,1)\)) and define the associated ordinary Dirichlet series \(D(s) = \sum a_n n^{-s}\). Then the system \((\varphi(n x))_{n}\) is a Riesz basis in \(L_2(0,1)\) if and only if \(D,1/D \in \mathcal{H}_\infty\), where \(\mathcal{H}_\infty\) denotes the Banach algebra of all ordinary Dirichlet series which converge pointwise to a bounded holomorphic function on \(\{\Re z >0\}\). In fact, this result marks the beginning of a modern (functional analysis oriented) theory of ordinary Dirichlet (see, e.g., the recent monographs [\textit{A. Defant} et al., Dirichlet series and holomorphic functions in high dimensions. Cambridge: Cambridge University Press (2019; Zbl 1460.30004)] and [\textit{H. Queffélec} and \textit{M. Queffélec}, Diophantine approximation and Dirichlet series. New Delhi: Hindustan Book Agency (2013; Zbl 1317.11001)]). A fundamental, independently interesting step in the proof of the above theorem shows that any given ordinary Dirichlet series \(D(s) = \sum a_n n^{-s}\) defines a bounded multiplier \(\mathcal{M}_D : \mathcal{H}_2 \to \mathcal{H}_2,\, E \mapsto D\ast E\) if and only if \(D \in \mathcal{H}_\infty\); here \(D \ast E\) is given by the Cauchy product and \(\mathcal{H}_2\) stands for the Hilbert space of all ordinary Dirichlet series with \(2\)-summable coefficients. The article under review shows a detailed analysis of these results. Recall that a system \((f_n)\) in a Hilbert space \(H\) is a frame whenever there are constants \(A,B >0\) such that for any finite sequence \(c=(c_n)\) of scalars we have that \(A \|c\|_{\ell_2} \leq \|\sum_n c_n f_n \|_H \leq B \|c\|_{\ell_2} \). If \((f_n)\) is even complete in \(H\), then it is said to be a Riesz basis, and it is a Riesz sequence whenever it is a Riesz basis of its closed span. Inspired by the work of Hedenmalm, Lindqvist, and Seip [loc.\,cit.]\, the main results of the present article characterize for a given system \((\varphi(n x))_{n}\) of dilations in \(L_2(0,1)\) (in terms of the associated Dirichlet series \(D\)), when the system satisfies an upper frame bound (i.e., is a Bessel sequence), when it satisfies a lower frame bound, and when it forms a Riesz sequence. Among other things, the proofs need `Bohr's vision' linking in a natural way the theory of ordinary Dirichlet series with harmonic analysis on the infinite dimensional circle group and infinite dimensional holomorphy on the unit ball of \(c_0\).