The conjectural bound for twisted incomplete Salié sums

Let aa and qq be integers with q2q\ge2, (a,q)=1(a,q)=1, and qq not a perfect square. Let H,K1H,K\ge1 and let λ,μ\lambda,\mu be real numbers. Write e(t)=e2πite(t)=e^{2\pi i t}, let h\overline h denote the inverse of hh modulo qq, and let (hq)\left(\frac{h}{q}\right) denote the relevant quadratic character. Twisted Salié-sum conjecture. For every ϵ>0\epsilon>0,

1hH(h,q)=1e(λh)0k<Ke(μk)(hq)e(ahk2q)ϵ(H1/2K1/2+H3/4+K+q1/2HK+q1/2K2)qϵ.\sum_{\substack{1\le h\le H\\(h,q)=1}}e(\lambda h)\sum_{0\le k<K}e(\mu k)\left(\frac{h}{q}\right)e\left(\frac{a\overline h k^2}{q}\right) \ll_\epsilon \left(H^{1/2}K^{1/2}+H^{3/4}+K+q^{-1/2}HK+q^{-1/2}K^2\right)q^\epsilon.

This estimate is introduced as a conjectural input for the paper's treatment of twisted incomplete Salié sums; its general validity is not established in the supplied text.

Sources & referencesView supporting material

Primary source

Tsz Ho Chan, “Distance between arithmetic progressions and perfect squares”, arXiv:1801.01605 (2018).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.