Elliptic logarithms, diophantine approximation and the Birch and Swinnerton-Dyer conjecture (Q2015847): Difference between revisions

Assuming the Birch and Swinnerton-Dyer conjecture (the precise version with the explicit constant), this paper proves an explicit upper bound for the \(x\)-coordinates of \(S\)-integral points on an elliptic curve (Theorem 3.4). The bound is in terms of the discriminant and the degree of the number field. As a consequence, it follows from the Chevalley-Weil theorem that the radical of \(abc\) has to be large when \(a+b=c\). The inequality they obtain is worse than the published results (e.g. [\textit{K. Györy}, Acta Arith. 133, No. 3, 281--295 (2008; Zbl 1188.11011)]), but this paper brings to light a connection between the BSD and the \(abc\) conjectures. To prove Theorem 3.4, the authors quote their earlier work [Ramanujan J. 32, No. 1, 125--141 (2013; Zbl 1362.11065)] bounding the \(S\)-integral points in terms of the regulator of the elliptic curve and the Néron-Tate heights of the generators of the Mordell-Weil group. This is where they use the known estimates of linear forms in elliptic logarithms ([\textit{S. David}, Mém. Soc. Math. Fr., Nouv. Sér. 62, 143 p. (1995; Zbl 0859.11048)]. The second author [``On the Mordell-Weil group and the Tate-Shafarevich group of an abelian variety'', Preprint, \url{arXiv:0801.1054}] has previously obtained an upper bound for the regulator in terms of the discriminant assuming BSD, and this in turn bounds the heights of the generators.