Explicit formulas for Dirichlet and Hecke \(L\)-functions (Q1765974)

In a former paper [J. Number Theory 65, 325--333 (1997; Zbl 0884.11036)] the author obtained a criterion for the validity of the Riemann Hypothesis. Let \[ \lambda_{n} = {1\over (n-1)!}{d^{n}\over ds^n }[s^{n-1}\log \xi(s)]_{s=1} \quad (n=1,2,3,\ldots) , \] where \[ \xi(s)=s(s-1)\pi^{-s/2}\Gamma \biggl({s\over 2}\biggr)\zeta(s) = \prod_{\rho} \biggl(1-{s\over \rho}\biggr) , \] the product being taken over all complex zeros \(\rho\) of the Riemann zeta-function \(\zeta(s)\) with \(\rho\) and \(1-\rho\) paired together. It is easy to see that \[ \lambda_{n} = \sum_{\rho} \biggl[1- \biggl(1-{1\over \rho}\biggr)^{n}\biggr], \] where the sum is understood as \(\sum_{\rho} = \lim_{T\to\infty}\sum_{| \Im(\rho)| \leq T}\). Li's criterion says the Riemann Hypothesis is equivalent to the statement that \(\lambda_{n}\geq 0\) for every positive integer \(n\). In the same paper the author also proved an identical criterion for the Dedekind zeta-function of a number field. Soon afterwards, \textit{E. Bombieri} and \textit{J.C. Lagarias} [J. Number Theory 77, 274--287 (1999; Zbl 0972.11079)] showed that Li's criterion holds for any multiset of complex numbers \(\rho\) satisfying the conditions: (i) \(0\) and \(1\) are not elements of the multiset; (ii) the multiset is invariant with respect to complex conjugation and the functional equation map \(s \mapsto 1-s\); (iii) \(\sum_{\rho}{1+| \Re(\rho)| \over (1+| \rho| )^2} < \infty\). Hence it emerged that Li's criterion is not specific to zeta functions. Bombieri and Lagarias also gave an arithmetic formula for the numbers \(\lambda_{n}\) via Weil's explicit formula, and provided an explanation that the positivity in Li's criterion has the same meaning as Weil's positivity condition for the Riemann Hypothesis. They remarked that the interesting point in Li's criterion is that it requires an explicit and rather simple set of test functions for its verification. We add that there are other criteria of this type, notably the one given by \textit{L. Aizenberg, V. Adamchik} and \textit{V. E. Levit} [Comput. Math. Appl. 37, No. 1, 87--94 (1999; Zbl 0937.11060)]. In this paper the author adapts the method of Bombieri and Lagarias to the case of Dirichlet \(L\)-functions and \(L\)-series associated to elliptic curves over rational numbers. In both cases, defining \(\lambda_{n}\) in the same form as the above sum over the nontrivial zeros, the arithmetic interpretation for the numbers \(\lambda_{n}\) are obtained by using Weil's explicit formula.