Upper Bounds for the Lowest First Zero in Families of Cuspidal Newforms

Xueyiming Tang,Steven J. Miller
2024-05-18
Abstract:Assuming the Generalized Riemann Hypothesis, the non-trivial zeros of $L$-functions lie on the critical line with the real part $1/2$. We find an upper bound of the lowest first zero in families of even cuspidal newforms of prime level tending to infinity. We obtain explicit bounds using the $n$-level densities and results towards the Katz-Sarnak density conjecture. We prove that as the level tends to infinity, there is at least one form with a normalized zero within $1/4$ of the average spacing. We also obtain the first-ever bounds on the percentage of forms in these families with a fixed number of zeros within a small distance near the central point.
Number Theory
What problem does this paper attempt to address?
The problem that this paper attempts to solve is, under the assumption that the Generalized Riemann Hypothesis (GRH) holds, for a family of even cusp newforms with the prime level tending to infinity, what is the upper bound of the position of the lowest first zero on the critical line. Specifically, the authors look for the position of the first non - trivial zero of these L - functions on the critical line and try to determine the specific bounds of this position relative to the average spacing. In addition, they also study the number and distribution of low - zeros near the central point of these L - functions, and in particular, give the upper bound of the proportion of forms in the family with a fixed number of zeros falling within a small interval near the central point. ### Main Results 1. **Upper Bound of the Lowest First Zero**: - By using the n - level central moment, the authors prove that as the level tends to infinity, there exists at least one form whose normalized zero distance from the central point is no more than 1/4 of the average spacing. - Specifically, for the support interval \(\sigma = 2\) and the test function \(h=\cos(\pi x / 2)\), the imaginary part \(\gamma\) of the first non - trivial zero of the even cusp newforms in this family falls within the interval \([- 0.25,0.25]\). 2. **Number and Distribution of Low - Zeros**: - The authors also give the upper bound of the proportion of forms in the family with a fixed number of zeros falling within a small interval near the central point. - For example, for a fixed \(\rho = 0.4\), when \(r = 6\), the probability upper bound given by the 2 - level density is 0.111085 (that is, at most about 11% of the forms have a zero within 40% of the average spacing near the central point), and the upper bound given by the 6 - level density is 1.58572 (more than 100%, so it is invalid). However, when \(r = 20\), the upper bound given by the 6 - level density is 0.0000121805, which is far better than 0.0042038 of the 2 - level density. ### Methods - **Construction of Test Function**: - The authors construct a test function \(\phi_\omega(x)\) such that it is non - negative within the interval \([-ω,ω]\) and non - positive within the interval \((-\infty,-ω)\cup(ω,\infty)\). - By choosing appropriate \(\sigma\) and \(\omega\), the properties of the test function can be optimized, thus obtaining more accurate results. - **n - level Central Moment**: - Use the n - level central moment to obtain the lower bound of the number of normalized zeros. - By analyzing these central moments, the authors are able to derive statistical results about the number and distribution of low - zeros. ### Conclusions - This paper successfully finds the upper bound of the position of the first non - trivial zero on the critical line of even cusp newforms by using the methods of n - level central moment and random matrix theory, and gives the upper bound of the proportion of forms in the family with a specific number of zeros falling within a small interval near the central point. - These results not only verify some predictions of the Generalized Riemann Hypothesis, but also provide new tools and methods for further research on the zero - distribution of L - functions.