Diophantine inequality involving binary forms (Q5919184)

From MaRDI portal
scientific article; zbMATH DE number 6409527
Language Label Description Also known as
English
Diophantine inequality involving binary forms
scientific article; zbMATH DE number 6409527

    Statements

    Diophantine inequality involving binary forms (English)
    0 references
    0 references
    27 February 2015
    0 references
    This paper is concerned with the study of Diophantine inequalities that are composed of binary forms. Let \(\phi_j\in \mathbb{Z}[x,y]\) for \(1\leq j\leq s\) be nondegenerate homogeneous forms of degree \(d=3\) or \(d=4\). Assume that \(\lambda_i\) for \(1\leq i\leq s\) are non-zero real numbers such that \(\frac{\lambda_1}{\lambda_2}\) is irrational, and the form \(\sum_{i=1}^s \lambda_i \phi_i(x_i,y_i)\) is indefinite. In this paper it is proved that under these assumptions, and for any \(\eta \in \mathbb{R}\) and \(\sigma< (2^{d-2}+1)^{-1}\) the Diophantine inequality \[ \left|\sum_{i=1}^s \lambda_i \phi_i(x_i,y_i) +\eta \right|< (\max_{1\leq i\leq s} \{|x_i|,|y_i|\})^{-\sigma}, \] has infinitely many integer solutions, as soon as \(s\geq 2^{d-1}+1\). The problem is motivated by work of \textit{T. D. Wooley} [Duke Math. J. 100, No. 3, 373--423 (1999; Zbl 1130.11312)] on Hua's lemma and exponential sums for binary forms. As before assume that \(\phi_j\in \mathbb{Z}[x,y]\) for \(1\leq j\leq s\) are homogeneous forms of degree \(d\) and non-zero discriminant. If \(s\) is sufficiently large (which means \(s>\max \{d^2, 2^{d-1}\}\) in the case \(d=3,4\)) and the form \(\sum_{i=1}^s \phi_i(x_i,y_i)\) is non-definite, then Wooley showed that the equation \(\sum_{i=1}^s \phi_i(x_i,y_i)=0\) has infinitely many integer solutions. More precisely, he obtained an asymptotic formula for the number of integer solutions. The proof of the main result of the paper under review is based on an application of the Davenport-Heilbronn circle method. It uses techniques and results from Wooley's work on binary forms. The proof crucially relies on estimates for the exponential sum \[ S(\alpha) = \sum_{x,y\sim X} e(\alpha \phi(x,y)), \] which are a form of Weyl's inequality and Hua's Lemma as stated in Theorem~1 and Theorem~2 in [Wooley, loc. cit.].
    0 references
    0 references
    Diophantine inequality
    0 references
    binary form
    0 references
    0 references