Geometric aspects of Lucas sequences. I (Q784830)

From the text: ``We present a way of viewing Lucas sequences in the framework of group scheme theory. This enables us to treat the Lucas sequences from a geometric and functorial viewpoint, which was suggested by \textit{R. R. Laxton} [Duke Math. J. 36, 721--736 (1969; Zbl 0226.10010)] ] and by \textit{M. Aoki} and \textit{Y. Sakai} [Rocky Mt. J. Math. 47, No. 8, 2513--2533 (2017; Zbl 1434.11039)].'' \dots ``In this article we study the divisibility problem for Lucas sequences from a geometric viewpoint, translating several descriptions on Lucas sequences into the language of affine group schemes. For example, let \(D\) denote the discriminant of the quadratic polynomial \(t^2 - Pt + Q\), that is to say, \(D = P^2 - 4Q\), and put \(A_D = \mathbb Z[t]/(t^2 - D)\) and \(G_{(D)} = \prod_{A_D/\mathbb Z} \mathbb G_{m,A_D} / \mathbb G_{m,\mathbb Z}\). Here \(\prod_{A_D/\mathbb Z}\) denotes the Weil restriction functor associated to the ring extension \(A_D/\mathbb Z\). In the Section 1, we introduce the affine group schemes denoted by \(\mathbb G_D\), \(U_D\) and \(G_{(D)}\), giving full explanation on the groups of \(\mathbb Q\)-rational points, \(\mathbb F_p\)-rational points and \(\mathbb Z_{(p)}\)-valued points. In the first half of Section 2, we give full explanation on \(G_D(\mathbb Z_p)\), \(U_D(\mathbb Z_p)\) and \(G_{(D)}(\mathbb Z_p)\) and deduce description on \(G_D(\mathbb Z_p/p^n\mathbb Z)\), \(U_D(\mathbb Z_p/p^n\mathbb Z)\) and \(G_{(D)}(\mathbb Z_p/p^n\mathbb Z)\) for \(n\ge 2\). In the latter half of Section 2, we define a homomorphism of group schemes \(p^n: G_{(p^{2n}D)} \rightarrow G_D\), which gives a description of the kernel of the reduction map \(G_D(\mathbb Z_p)\rightarrow G_{(D)}(\mathbb Z/p^n\mathbb Z)\). In the first half of Section 3, after relating Lucas sequences with the group schemes \(G_D\) and \(G_{(D)}\), we present an interpretation on the notion of rank and period for Lucas sequences in our context. Moreover, we give new proofs for more or less known facts, some of which go back to Lucas, Carmichael and Lehmer, for example. In the latter half of Section 3, we reformulate and generalize remarkable results of Aoki-Sakai [loc. cit.], which suggests a way to treat Lucas sequences geometrically. In the Section 4, we reconstruct the theory developed by Laxton [loc. cit.; Pac. J. Math. 32, 173--179 (1970; Zbl 0226.10011)], who defined an interesting group \(G(f)\) so that the divisibility problem for Lucas sequences might be dealt with systematically. For example, we give an explicit description of the group \(G(f)\) as follows: Theorem 4.2. Let \(P\) and \(Q\) be non-zero integers with \((P,Q) = 1\), and put \(f(t) = t^2 - Pt + Q\). Let \(p\) be an odd prime. Then there exists a canonical isomorphism \[ \omega: G_{(D)}(\mathbb Q)/\Theta\rightarrow G(f). \] Laxton's work is pioneering, but seems unfortunately ignored and forgotten. A main reason may be that Laxton did not give an explicit description of \(G(f)\). It would be surprising that we can describe the groups \(G(f)\), \(H(f,p)\), \(K(f,p)\) and \(G(f, p^n)\) \((n\ge 1)\) defined by Laxton through the groups \(G_{(D})(\mathbb Q) = U_D(\mathbb Q)\), \(U_D(\mathbb Z(p))\), \(G_{(D)}(\mathbb Z(p))\) and \(G_{(p^{2n}D)}(\mathbb Z(p))\) \((n\ge 1)\), as is shown in Theorem 4.2 and Corollaries 4.3 and 4.9. The last three subsections form the coda of the Section 4. We summarize there argument on the divisibility problem for Lucas sequences, recalling the related results established by \textit{M. Ward} [Ill. J. Math. 6, 40--52 (1962; Zbl 0099.02704)] and Laxton [loc. cit.]. We conclude the introduction by referring to a related work by Ward [loc. cit.]. He investigated Lucas sequences defined for \(p\)-adic integers, using systematically the \(p\)-adic logarithmic function. His method is essentially equivalent to the ones employed in the Section 2.'' An interesting paper reviving classical texts of Laxton and Ward in a new environment and the reviewer is looking forward for Part II [Tokyo J. Math. 43, No. 2, 383--454 (2020; Zbl 1484.11065)]. The corrigendum corrects mistakes in the statements of Corollary 3.13 and Corollary 3.14.