Integral zeros of a polynomial with linear recurrences as coefficients (Q2020433)

Let \(K\) be a number field, let \(S\) a finite set of absolute values on \(K\) and denote by \(\mathcal O_S\) the set of \(S\) integers of \(K\). Let \(G_n^{(i)}, i=0,1,\ldots,d\) be linear recurrence sequences taking integer vales for any \(n\). For \(n\in \mathbb N\) consier the roots \(z\in\ \mathcal O_S\) of the equation \[ G_n^{(0)}Z^d+\cdots G_n^{(d-1)}Z+G_n^{(d)}=0. \] In other words consider the solutions in \((n,z)\in \mathbb N\times \mathcal O_S\) of this equation.\par Using the Binet representation and the characteristic roots of the linear recurrence sequences the equation can be written in the form \[ a_0(\beta_1^n,\ldots,\beta_r^n)Z^d+\cdots + a_{d-1}(\beta_1^n,\ldots,\beta_r^n)Z+a_d(\beta_1^n,\ldots,\beta_r^n)=0 \] with linear polynomials \(a_0(X_1,\ldots,X_r),\ldots,a_d(X_1,\ldots,X_r)\). Hence the problem translates into a polynomial-exponentian diophantine equation. Under certain technical assumptions the authors descibe the set os zeros \((n,z)\in \mathbb N\times \mathcal O_S\) of the equation. They also prove a statement in the spirit of Hilbert's irreducibility theorem for such polynomials. \textit{C. Fuchs} and \textit{A. Scremin} [Publ. Math. 65, No. 1--2, 149--172 (2004; Zbl 1064.11031)] considered integral solutions of similar equations viewing them as polynomial-exponential equations applying [\textit{P. Corvaja} and \textit{U. Zannier}, Compos. Math. 131, No. 3, 319--340 (2002; Zbl 1010.11038)]. Similar results can be found in [\textit{P. Corvaja} and \textit{U. Zannier}, Acta Arith. 94, No. 1, 25--40 (2000; Zbl 0963.11020)]. The proof is based on the method of \textit{P. Corvaja} and \textit{U. Zannier}, see e.g. [Applications of Diophantine approximation to integral points and transcendence. Cambridge: Cambridge University Press (2018; Zbl 1452.11004)].