No label defined (Q4533267)

For integral \(m\) with \(m\geq 2\) put \[ R_m(x)=x(x+1)\cdots(x+m-1)\quad \text{and}\quad S_m(x)=1^m+2^m+\hdots+(x-1)^m. \] The literature of Diophantine equations involving any of the polynomials \(R_m(x)\) and \(S_m(x)\) is very rich. Such equations have been studied by Beukers, Brindza, Győry, Pintér, Saradha, Shorey, Tijdeman and many others. In the present paper the authors prove two interesting and compact results. On the one hand, they show that if \(m\geq 2\) and \(n\geq 1\), \((m,n)\neq (2,1)\), then the equation \(R_m(x) = S_n(y)\) has at most finitely many solutions in integers \(x,y\). On the other hand, it is also proved that if \(n>m\geq 1\), then the equation \(S_m(x)=S_n(y)\) has at most finitely many solutions in integers \(x,y\). To prove their theorems, the authors use a recent, very deep and general result on the Diophantine equation \(f(x)=g(y)\) due to \textit{Yu. F. Bilu} and \textit{R. F. Tichy} [ Acta Arith. 95, 261-288 (2000; Zbl 0958.11049)]. As this result is of an ineffective nature, the above two theorems are also not effective. However, in the case when \(n=1,3\) or \(m=2,4\), the authors apply Baker's theory to obtain effective bounds for the integer solutions of the equation \(R_m(x)=S_n(y)\).