Joint discrete universality of Dirichlet \(L\)-functions (Q2514371): Difference between revisions

We quote the authors' abstract: ``In this paper we prove a generalized version of a joint discrete universality theorem on the approximation of a collection of analytic functions by discrete shifts of Dirichlet L-functions.'' More in the detail, they give this kind of result (their Theorem 1.4), under a hypothesis of linear independence over the rational numbers of the \(r\) numbers \(h_j>0\) (\(j=1,\ldots,r\)); namely, the set \(L(h_1,\ldots,h_r;\pi)\) made of the number \(\pi\) and the numbers: \(h_j\log p\), for all \(j\leq r\) and for all primes \(p\), is linearly independent over \(\mathbb{Q}\). The shifts are taken as integer multiples of these \(h_j\), namely the shifts are \(kh_j\), with \(k>0\) integers. Also, the authors remark that they do not know of any specific instance of the positive numbers \(h_1,\ldots,h_r\) such that the relative set \(L(h_1,\ldots,h_r;\pi)\) is linearly independent over the rationals, but they give their Proposition 1.5, proving that the set of positive real numbers \(h_1,\ldots,h_r\) which do not satisfy this property has Lebesgue measure zero. The roof is short and elegant. As regards the proof of their main result (Theorem 1.4 quoted above), it is conducted by the usual blend of results, in the environment of analytic number theory, from ergodic theory, probability theory (Haar measures and weak convergence theorems) and approximation of analytic functions (Mergelyan Theorem).