Centralized variant of the Li criterion on functions fields (Q2662042)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Centralized variant of the Li criterion on functions fields |
scientific article |
Statements
Centralized variant of the Li criterion on functions fields (English)
0 references
8 April 2021
0 references
Let \(K\) be a function field over a finite field \(\mathbb F_q\) and \(X\) be its smooth projective curve of genus \(g\) over \(\mathbb F_q\). The zeta function of \(K\) is expressed as \[ Z_K(T)=\frac{L(T)}{(1-T)(1-qT)}, \] where \[ L(T)=\prod_{j=1}^{2g}(1-\alpha_j T)\in\mathbb Z[T] \] is a polynomial of degree \(2g\). We order the indices \(j\in\{1,\ldots,g\}\) so that \(\alpha_{g+j}=\overline{\alpha_j}\) and put \(\theta_j\) as \(\alpha_j=q^{1/2}\exp(i\theta_j)\). The completed zeta function is defined by \[ \xi_K(s)=q^{gs}L(q^{-s}). \] The Voros-Li coefficients \(\lambda_{K,\omega}^V(n)\) are defined by \[ \log\xi_K\left(\frac12+\frac{\sqrt{\omega s}}{1-s}\right) -\log\xi_K(1/2)=\sum_{n=1}^\infty\frac{\lambda_{K,\omega}^V(n)}n s^n. \] Assume \(\xi_K(1/2)\ne0\). The first main theorem in this article asserts the equivalence of GRH for \(\xi_K\) and the inequalities \[ \lambda_{K,\omega}^V(n)\ge0\quad(\text{ for all } n\ge 1) \] for \(0<\omega<4\min_j(\theta_j/\log q)^2\). The authors also introduce another type of Li-type coefficients which they call the Sekatskii-Voros-Li coefficients \(\lambda_{K,l}^{SV}(n)\) defined by \[ \lambda_{K,l}^{SV}(n) =\frac1{(2n-1)!}\frac{d^{2n}}{ds^{2n}}[L_{2n,l}(s)\log\xi_K(s)]_{s=1/2}, \] where \[ L_{2n,l}(s) =4\sum_{k=0}^{n-1}(n-k)A_{k,l,n}(s-1/2)^{2k} \] and \[ A_{k,l,n}=l^{2k-2n}\sum_{m=k}^n\binom {2n}{2m}\binom m k. \] The second main theorem shows that the GRH for \(\xi_K\) is equivalent to the non-negativity \[ \lambda_{K,l}^{SV}(n)\ge0\quad(\text{ for all } n\ge 1) \] with \(l>\max_j\left(\frac{\log q}{\theta_j}\right)\). As applications they deduce formulas and give observations for the case of \(n=1\) and \(n=2\) by using GRH proved to be true over \(K\).
0 references
function fields
0 references
Li's criterion
0 references
approximation of \(L\)-functions
0 references
0 references
0 references
0 references