The multiplicity of binary recurrences (Q1360895): Difference between revisions

Let \((u_m)_{m \in\mathbb{Z}}\) be a complex linear recurrence sequence. For any complex number \(c\) one defines \(U(c)\) to be the number of indices \(m\in \mathbb{Z}\) for which \(u_m =c\). The multiplicity of the recurrence sequence is defined to be \(U= \sup_c U(c)\). The sequence is called degenerate if its companion polynomial has at least a pair of distinct roots whose ratio is a root of unity; otherwise, it is called non-degenerate. A longstanding conjecture is that an arbitrary complex non-degenerate recurrence sequence of order \(n\geq 2\) has multiplicity bounded by a function that depends only on \(n\). In the paper under review, the author proves the conjecture for \(n=2\) by showing that, in this case, \(U\leq 2^{2^{23}}\). In his proof he has to make one further very natural assumption; namely, that at least one root of the second degree companion polynomial is not a root of unity. For ternary sequences he proves that \(U(0)\leq 2^{2^{23}}\). Two basic tools of the proof are a quantitative sharpened version of the subspace theorem in dimension 2 and a result by the author and \textit{E. Wirsing}, which gives universal lower bounds for the height of solutions in algebraic numbers of linear equations \(ax+by +1=0\) with complex coefficients [Lower bounds for the heights of solutions of linear equations, Invent. Math. 129, 1-10 (1997; Zbl 0883.11013)].