Isometries between spaces of multiple Dirichlet series (Q2414768): Difference between revisions

This interesting article sets the ground of a systematic structure theory of \(k\)-multiple Dirichlet series \(D=\sum_{m_1, \dots, m_k=1}^\infty \frac{a_{m_1, \dots, m_k}}{m_1^{s_1}\cdots m_1^{s_k}}\), where the \(a_{m_1, \dots, m_k} \in \mathbb{C}\) form the coefficients and the \(s_1, \dots, s_k \in \mathbb{C}\) are complex variables. This is done in the spirit of fundamental work on ordinary Dirichlet series due to Bohr, Hardy, and Riesz, among others, and the very recent renaissance of these ideas (see, e.g., the monographs [the reviewer 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)]). In a first stage, the authors clarify that they consider so-called regular convergence of \(k\)-multiple series \(\sum_{m_1, \dots, m_k=1}^\infty b_{m_1, \dots, m_k}\), i.e., for all \(j\) all subseries, formed by the sum over \(j\) many indexes \(m_{i_1}, \dots, m_{i_j}\), fixing all other indexes, converge. As in the ordinary case \(k=1\), it is proved that each \(k\)-multiple Dirichlet series \(D\) defines a largest \(k\)-multiple half-plane \([Re >\sigma_1] \times \dots \times [Re >\sigma_k]\) on which it converges regularly pointwise to a holomorphic function. Denote by \(\mathcal{H}_\infty ([Re >0]^k)\) the linear space of all \(k\)-multiple Dirichlet series \(D\) which on \([Re >0]^k\) converge regularly pointwise to a bounded (and then necessarily holomorphic) function. The authors show that \(\mathcal{H}_\infty ([Re >0]^k)\) (endowed with the supremum norm and the \(k\)-multiple Cauchy product) forms a Banach algebra. The proof combines (a quantified version of) a fundamental fact due to Bohr showing that, for ordinary Dirichlet series, the abscissas of boundedness and uniform convergence coincide, with a clever trick on vector-valued \(k\)-multiple Dirichlet series \(D\). As the main result, the authors then prove that \(\mathcal{H}_\infty ([Re >0]^k)\) may be identified isometrically with \(H_\infty(B_{c_0^k})\), the Banach space of all holomorphic functions on the \(k\)th Cartesian product \(c_0^k\) of the Banach space \(c_0\) of all null sequences; here the monomial coefficients and the Dirichlet coeffcients are preserved in the sense that for each choice of \(k\) many multi indices \(\alpha_j\) we have \(\frac{\partial^{(\alpha_1, \dots,\alpha_k)} f (0)}{\alpha_1! \cdots \alpha_k!} = a_{\mathfrak{p}^{\alpha_1}, \dots, \mathfrak{p}^{\alpha_k}}\), where \(\mathfrak{p}\) stands for the sequence of prime numbers. Obviously, \(c_0\) and \(c_0^k\) may be identified, and hence, as a consequence, the authors derive the interesting fact that the Banach algebras \(\mathcal{H}_\infty ([Re >0]^k)\) are all isometrically isomorphic.