New estimates of double trigonometric sums with exponential functions (Q2502310)

The authors give a new bound for the exponential sum \[ W_a(\gamma; T; \mathcal X; \mathcal Y)=\sum_{x\in {\mathcal X}} \bigg| \sum_{y \in {\mathcal Y}} \gamma(y) \exp(2\pi i a \lambda^{xy}/p)\bigg|, \] where \(a\) is an integer coprime to \(p\), \(\lambda\) is an element of the residue ring modulo a large prime number \(p\), \(\mathcal X\) and \(\mathcal Y\) are arbitrary subsets of the residue ring modulo \(p - 1\) and \(\gamma (n)\) are any complex numbers with \(| \gamma (n)| \leq 1\). Recently, in [Int. Math. Res. Not. 2005, No. 17, 1005--1014 (2005; Zbl 1070.11032)] the first author suggested a new approach to investigate the above sum. This approach allows to improve a series of previously known bounds. The aim of the present paper is to modify the argument of the cited paper and improve the upper bound for \(W_a(\gamma; T; \mathcal X; \mathcal Y)\) as follows. Theorem 1. For any given positive integer \(k\), the following estimate holds: \[ W_a(\gamma; T; \mathcal X; \mathcal Y)\ll \frac{|\mathcal X|^{1-1/2k} |\mathcal Y|^{1-1/(2k+2)} p^{1/2k+ 3/(4k+4)+o(1)}}{T^{1/(2k+2)}}, \] where the implied constants in the Vinogradov symbol \(\ll\) depends on the small positive quantity \(\varepsilon\) and on the fixed positive integer \(k\).