On probabilistic generalizations of the Nyman-Beurling criterion for the zeta Function (Q2071239)

Let \(\{t\}\) denote the fractional part of the real number \(t\), and \(\chi\) the indicator function of \((0,1)\). \textit{B. Nyman} [On the one-dimensional translation group and semi-group in certain function spaces. Uppsala: Appelbergs Boktryckeri (Diss.) (1950; Zbl 0037.35401)] proved in 1950 that the Riemann hypothesis (RH) holds true if and only if the quantity \[ \int_0^{\infty} \big(\chi(t)-\sum_{k=1}^nc_k\{\theta_k/t\}\big)^2 \, dt \] can be made arbitrarily small, choosing a suitable integer n, and real numbers \(c_k\) and \(\theta_k\) such that \(0<\theta_k \le 1\). In 2003, \textit{L. Báez-Duarte} [Atti Accad. Naz. Lincei, Cl. Sci. Fis. Mat. Nat., IX. Ser., Rend. Lincei, Mat. Appl. 14, No. 1, 5--11 (2003; Zbl 1097.11041)] strengthened this criterion by showing that it is enough to take \(\theta_k=1/k\). In this well-written article, the authors give probabilistic extensions of these criteria, by taking the \(\theta_k\) as random. Four theorems are proved, among which the following is typical (Theorem 2.6). Let \((Y_{k,n})_{1\le k\le n}\) be an array of nonnegative, square-integrable random variables, and \(X_{k,n}=Y_{k,n}^2\). Suppose that \begin{align*} \mathbb{E}\,Y_{k,n}&=\frac{1}{\sqrt{k}} \quad (1\le k \le n)\\ \sup_{1\le k\le n} \mathbb{V}\, Y_{k,n}&\ll n^{-3-\theta}\\ \mathbb{P}\,(Y_{1,n} \ge 1) & \le 1-\nu, \end{align*} for some \(\theta >0\) and \(\nu \in (0,1)\). Then (RH) holds if and only if \[ \mathbb{E}\,\int_0^{\infty} \big(\chi(t)-\sum_{k=1}^nc_{k,n}\{X_{k,n}/t\}\big)^2 \, dt \rightarrow 0 \quad (n \rightarrow \infty), \] for a suitable array \((c_{k,n})_{1\le k \le n}\) of real numbers.