On the exponential Diophantine equation \(F_{n+1}^x - F_{n-1}^x = F_m^y\) (Q2167797): Difference between revisions

Let \( (F_n)_{n\ge 0} \) be the usual sequence of Fibonacci number defined by the linear recurrence: \( F_0=0, ~ F_1=1, \) and \( F_{n+2}=F_{n+1}+F_n \) for all \( n\ge 0 \). In the paper under review, the authors study the exponential Diophantine equation: \[ F_{n+1}^x-F_{n-1}^x=F_m^{y},\tag{1} \] in positive integers \(n, m, x\) and \(y\). This is an extension of a previous problem considered by \textit{B. K. Patel} and \textit{A. P. Chaves} in [Mediterr. J. Math. 18, No. 5, Paper No. 187, 11 p. (2021; Zbl 1490.11042)], by adding an extra exponent ``\( y \)''. In their main result, the authors prove that the exponential Diophantine equation (1) does not have any other solutions in positive integers \( (n,m,x,y) \), except the trivial ones: \[ (n,m,x,y) \in \{(1,1,x,y), (1,2,x,y), (2,1,1,y), (2,2,1,y): x,y\ge 1\} \cup \{(3,3,2,3), (6,3,1,2)\}, \] as well as the parametric ones: \( (n,m,x,y)\in \{(n,n,1,1), ~(n,2n,2,1):~n\ge 1\} \). The proof of their main result follows from a clever combination of techniques in Diophantine number theory, the usual properties of the Fibonacci numbers, the Baker's theory of nonzero linear forms in logarithms of algebraic numbers, as well as the reduction techniques involving the theory of continued fractions plus the LLL-algorithm. The rigorous computations are carefully done with the aid of a computer program in \texttt{Mathematica}.