scientific article; zbMATH DE number 1130562

From MaRDI portal
Publication:4382429





Let \(a\), \(b\) and \(k\) be fixed integers with \(k \geq 3\). The authors' result is a strong quantitative form of the proposition that almost all of the integers \(n\) represented by the binary form \(ax^k-by^k\) are so representable in an essentially unique way (i.e., in a way not differing in other than the sign or order of \(x\), \(y\)). Let \(\mu(X)\) denote the exceptional set of the number of these \(n\) with \(|n|\leq X\) that have multiple representations, so that the trivial estimate would read \(\mu(X) \ll X^{2/k}\). The authors' result is \(\mu(X) \ll X^{c/k+\epsilon}\), for a certain explicitly given \(c=c(k)<2\) with \(c \sim {3 \over 2}\) as \(k \to \infty\). Thus they can estimate the total number of \(n\) with \(|n|\leq X\) represented by the form, by an expression incorporating an error term \(O( X^{c/k})\). In the case when \(a = \pm b\) there are results already in the literature. Using a sieve method and Deligne's estimates for exponential sums \textit{C. Hooley} obtained a result with \(c \sim {5 \over 3}\) when the degree \(k\) is odd [Proc. Lond. Math. Soc., III. Ser. 43, 73-109 (1981; Zbl 0463.10035)], which he has more recently extended to even \(k\) in the case \(a=b\) [J. Reine Angew. Math. 475, 55-75 (1996; Zbl 0848.11041)]. Actually these results have \(c = {5 \over 3}\) if \(k=3\), or if \(a=b\). Furthermore when \(k=3\) \textit{D. R. Heath-Brown} [Acta Arith. 79, No. 1, 17-30 (1997; Zbl 0863.11021)] has improved the result to \(c={4 \over 3}\), by a development of an elementary approach due to Wooley. Hooley has also obtained weaker results for more-or-less arbitrary forms of degree three, and for a wide class of forms of degree four. The present authors' result is the first to apply when \(|a|\neq |b|\) and when \(k \geq 5\). They employ a ``slicing approach previously used by \textit{C. M. Skinner} and \textit{T. D. Wooley} in the case \(a=b\) [J. Reine Angew. Math. 462, 57-68 (1995; Zbl 0820.11059)]. In this approach, the problem is reduced to one concerning the number of points on a carefully selected collection of plane curves, which are then attacked using a theorem of \textit{E. Bombieri} and \textit{J. Pila} [Duke Math. J. 59, No. 2, 337-357 (1989; Zbl 0718.11048)]. This approach raises the possibly serious difficulty that these curves, of degree \(k\), have to be shown to be absolutely irreducible. One approach to this question is to try to show that they have no singular points. The authors' method depends on showing rather less, viz. that they have less than \(k-2\) singular points.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4382429)