On a question of Heath-Brown: Good estimates on average for higher degree Kloosterman sums (Q4807669)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 1916205
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | On a question of Heath-Brown: Good estimates on average for higher degree Kloosterman sums |
scientific article; zbMATH DE number 1916205 |
Statements
On a question of Heath-Brown: Good estimates on average for higher degree Kloosterman sums (English)
0 references
25 May 2003
0 references
higher degree form
0 references
averages
0 references
complete exponential sums
0 references
composite modulus
0 references
0.8392311930656433
0 references
0.8383032083511353
0 references
0.8281804919242859
0 references
0.8171723484992981
0 references
0.8149004578590393
0 references
This paper is concerned with complete exponential sums \[ S(F(x_1,\dots, x_n);q)= \sum_{1\leq x_1,\dots, x_n\leq q}\exp (2\pi iF(x_1,\dots, x_n)/q). \] When \(q\) is prime and \(F\) is reasonably well behaved these sums can be estimated efficiently via Deligne's Riemann Hypothesis. One can then handle square-free values of \(q\) by an elementary multiplicative property of the sums \(S(F;q)\). However the case in which \(q\) is a prime power, or a general composite modulus, is considerably harder, and is properly understood only when \(n=1\). In this paper one takes \(F\) to be a fixed form of degree \(d\geq 2\) and defines \[ T(b_1,\dots, b_n;q)= \sum_{1\leq a\leq q, (a,q)=1} S(aF(x_1,\dots, x_n)+ b_1x_1+\cdots+ b_nx_n;q) \] and \[ {\mathcal S}_q(B)= q^{-n} \mathop{{\sum}^*}_{-B\leq b_1,\dots, b_n\leq B} |T(b_1,\dots, b_n;q)|, \] where \(\sum^*\) denotes the omission of the \(n\)-tuple \((0,\dots, 0)\). Averages of this type have been important in circle method applications. Thus in the case \(d=3\) the reviewer [Proc. Lond. Math. Soc. (3) 47, 225-257 (1983; Zbl 0494.10012)] encountered the sum \({\mathcal S}_q(B)\) in his treatment of the Hasse principle for nonsingular cubic forms in 10 variables, and showed that there is a constant \(C= C(F)\) such that \[ S_q(B)\leq C^{\omega(q)} (\log (q+1))^{2n} q^{1-n/2} B^n, \quad B\geq q^{1/3}. \] Here \(\omega(q)\) denotes the number of distinct prime factors of \(q\). Thus \[ S(aF(x_1,\dots, x_n)+ b_1x_1+\cdots+ b_nx_n;q) \] has average size essentially of order \(q^{n/2}\), which is optimal. The principal result of the present paper applies to a nonsingular form of arbitrary degree \(d\), and shows that \[ S_q(B)\leq C^{\omega(q)} q^{1-n/2} B^n, \quad B\geq q^{1/2}. \] While it would be nice to have such a result for smaller values of \(B\), it is a considerable achievement to produce such a bound for arbitrary degrees. It should be stressed that the main difficulty in proving this result lies in the case \(q=p^r\) with \(r\geq 2\). The further elucidation of multiple exponential sums to prime power modulus remains a difficult and important goal.
0 references