Computation of the zeros of the \(L\)-function associated with the cubic theta function (Q843565): Difference between revisions

The paper is devoted to the computational results for the zeros of the \(L\)-function associated with the cubic theta-function. The function \(L(\tau;s)\), \(s=\sigma+it\), is defined by the Dirichlet series \[ L(\tau;s)=\sum_{\nu}\frac{\tau(\nu)}{||\nu||^s}, \quad \text{Re}\, s>1, \] where \(\tau(\nu)\) are the Fourier coefficients of the cubic theta function, \(||z||\) is the norm mapping \(\mathbb Q(\sqrt{-3})\to \mathbb Q\), and, for \(z \in \mathbb Q(\sqrt{-3})\), \(||z||=z \cdot {\overline z}\). For the numerical results, the author uses conjecture that the nontrivial zeros of the function \(L(\tau;s)\) are simple, and they are located in the critical strip \(\{s \in \mathbb C: 0< \text{Re}\, s <1\}\). One of computational result shows that, for not large \(T\), the ratio of number of zeros \(E(T)\) on the critical line and number of nontrivial zeros \(N(T)\) with imaginary parts ranging from 0 to \(T\) first strongly varies and then slowly decreases, and finally its almost stabilizes. The author conjectures, that for sufficiently large \(T\), \(\gamma<\frac{E(T)}{N(T)}<\delta\) for some \(\gamma,\delta \in \mathbb R\), \(0<\gamma<\delta<1\), and, moreover, that the limit \(\frac{E(T)}{N(T)}\) exists and belongs to the interval \((0,1)\) as \(T \to \infty\). Second computational result deals with the density of zeros of \(L(\tau;s)\) located in quadrant \(\text{Re}\,s>\frac{1}{2}\), \(\text{Im}\, s \geq 0\).