Equivalents of the Riemann hypothesis (Q2492893)

This interesting paper deals with the value-distribution of the Riemann zeta-function \(\zeta(s)\). The authors obtain necessary and sufficient conditions for the truth of the Riemann hypothesis for \(\zeta(s)\) in terms of the functional distribution of Dirichlet \(L\)-functions \(L(s,\chi_d)\) in the right half of the critical strip; here \(\chi_d\) denotes the quadratic character \(\chi_d\bmod\,| d|\) given by the Kronecker symbol. They show that the Riemann hypothesis is true for \(\zeta(s)\) if and only if, for any compact subset \(K\) of the open strip \(1/2<\text{Re}\,s<1\) (satisfying a certain necessary topological condition) and any positive \(\varepsilon\), \[ \liminf_{X\to\infty}{1\over D(x)}\sharp\{1\leq d\leq X: \max_{s\in K}| L(s,\chi_d)+ \zeta(s)|<\varepsilon\}>0, \] where \(D(x)\) is the number of fundamental discriminants \(d\leq x\); an analogous result is stated and proved with respect to prime discriminants \(d\). The proofs of these results make use of Rouché's theorem and a zero-density estimate for quadratic \(L\)-functions; however, the most important ingredient is a universality theorem for families of quadratic \(L\)-functions from the authors [Trans. Am. Math. Soc. 358, No. 10, 4343--4366 (2006; Zbl 1160.11041), Acta Arith. 123, 143--161 (2006)]. The idea of linking approximating properties of Dirichlet series with the Riemann hypothesis dates back to Bagchi and Harald Bohr (in the language of almost-periodicity).