Representations of integers by systems of three quadratic forms

DOI10.1112/PLMS/PDW027MaRDI QIDQ2831998zbMATH OpenOpenAlexFDO

Authors Damaris Schindler, Melanie Matchett Wood, Lillian Beatrix Pierce

Publication date 4 November 2016

Published in Proceedings of the London Mathematical Society. Third Series (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1509.04757

geometry Kloosterman refinement quadratic forms representation simultaneous system three forms

Counting solutions of Diophantine equations (11D45) Representation problems (11D85) Quadratic forms over global rings and fields (11E12) Applications of the Hardy-Littlewood method (11P55)

Abstract: It is classically known that the circle method produces an asymptotic for the number of representations of a tuple of integers

(n_{1}, l d o t s, n_{R})

by a system of quadratic forms

Q_{1}, l d o t s, Q_{R}

in

k

variables, as long as

k

is sufficiently large; reducing the required number of variables remains a significant open problem. In this work, we consider the case of 3 forms and improve on the classical result by reducing the number of required variables to

k g e q 10

for "almost all" tuples, under appropriate nonsingularity assumptions on the forms

Q_{1}, Q_{2}, Q_{3}

. To accomplish this, we develop a three-dimensional analogue of Kloosterman's circle method, in particular capitalizing on geometric properties of appropriate systems of three quadratic forms.

Review text

Let \(Q_1(\mathbf{x}),Q_2(\mathbf{x}),Q_3(\mathbf{x})\) be integral quadratic forms in \(k\) variables. This paper concerns the simultaneous representation problem \(Q_i(\mathbf{x})=n_i\), for \(i=1,2,3\). Write \(R_B(\underline{n})\) for the number of solutions with \(|\mathbf{x}|\leq B\). Under suitable conditions one would hope for an asymptotic formula NEWLINE\[NEWLINER_B(\underline{n})=CB^{k-6}+o(B^{k-6})\tag{*}NEWLINE\]NEWLINE when \(B\rightarrow\infty\), where \(C\) is the product of the standard singular integral and singular series. For example, when the forms \(Q_1,Q_2,Q_3\) satisfy the Jacobian nonsingularity condition the method of \textit{B. J. Birch} [Proc. R. Soc. Lond., Ser. A 265, 245--263 (1962; Zbl 0103.03102)] shows that (*) holds as soon as \(k\geq 27\).NEWLINENEWLINEIn order to handle fewer variables the paper asks what happens for ``almost all values \(n_1,n_2,n_3\leq B^2\). Birch's method would establish (*) for almost all such \(\underline{n}\) if \(k\geq 14\). The principal result of the paper is a significant improvement on this, requiring only that \(k\geq 10\). The method used is a development of that described by the first author in joint work with the reviewer [J. Reine Angew. Math. 727, 85--143 (2017; Zbl 1396.11121)], which treats pairs of quadratics, using a two dimensional Kloosterman refinement. When one passes from pairs of forms to triples there are significant structural changes needed for the argument. In particular the simple Jacobian smoothness condition is not sufficient. Instead one considers the ternary form NEWLINE\[NEWLINEF(x,y,z)=\det(xQ_1+yQ_2+zQ_3)NEWLINE\]NEWLINE and requires that \(F\) should have non-zero discriminant. The second section of the paper gives an informative description of the geometry of intersections of 3 quadrics, highlighting the extra difficulties which arise compared with the situation for pairs of quadrics.NEWLINENEWLINEOne might ask what happens for 4 or more quadrics, but in this case the form analogous to \(F\) vanishes identically, so that a different approach will be required.

Recommendations

Cites work

Cited in

(13)

This page was built for publication: Representations of integers by systems of three quadratic forms

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2831998)