English

Evaluation of exponential sums and Riemann zeta function on quantum computer

Quantum Physics 2020-02-26 v1 Mathematical Physics math.MP Number Theory

Abstract

We show that exponential sums (ES) of the form \begin{equation*} S(f, N)= \sum_{k=0}^{N-1} \sqrt{w_k} e^{2 \pi i f(k)}, \end{equation*} can be efficiently carried out with a quantum computer (QC). Here NN can be exponentially large, wkw_k are real numbers such that sum Sw(M)=k=0M1wkS_w(M)=\sum_{k=0}^{M-1} w_k can be calculated in a closed form for any MM, Sw(N)=1S_w(N)=1 and f(x)f(x) is a real function, that is assumed to be easily implementable on a QC. As an application of the technique, we show that Riemann zeta (RZ) function, ζ(σ+it)\zeta(\sigma+ i t) in the critical strip, {0σ<1,tR}\{0 \le \sigma <1, t \in \mathbb{R} \}, can be obtained in polyLog(t) time. In another setting, we show that RZ function can be obtained with a scaling t1/Dt^{1/D}, where D2D \ge 2 is any integer. These methods provide a vast improvement over the best known classical algorithms; best of which is known to scale as t4/13t^{4/13}. We present alternative methods to find S(f,N)\lvert S(f,N) \rvert on a QC directly. This method relies on finding the magnitude A=0N1akA=\lvert \sum_0^{N-1} a_k \rvert of a nn-qubit quantum state with aka_k as amplitudes in the computational basis. We present two different ways to do obtain AA. Finally, a brief discussion of phase/amplitude estimation methods is presented.

Cite

@article{arxiv.2002.11094,
  title  = {Evaluation of exponential sums and Riemann zeta function on quantum computer},
  author = {Sandeep Tyagi},
  journal= {arXiv preprint arXiv:2002.11094},
  year   = {2020}
}
R2 v1 2026-06-23T13:53:37.626Z