Shifted power of a polynomial with integral roots (Q6175360)

In 2013, A. Schinzel formulated a problem due to W. Bednarek: for what positive integers \(n\) there exist monic polynomials \(f(x)\in \mathbb Z [x] \) with \(deg f=n\) such that \(f(x)^2-1\) has exactly \(2n\) integral roots (counting their multiplicities). This problem was negatively solved recently by \textit{A. Łydka} [Math. Slovaca 71, No. 4, 821--822 (2021; Zbl 1487.11030)] for \(n\geq 3\) (for \(n=1 \) and \(n=2\), the polynomials \(f(x)=x\) and \(f(x)=x^2+x-1\) respectively are obvious solutions). In this paper, the author solved analogous problem for polynomials of the form \(f(x)^m-c, c\in \mathbb Z, c\geq 1, m, n\geq2\). The main result, Theorem 1.1, states that \(f(x)^m-c\) has \(mn\) roots (counting with multiplicities) if and only if \(m=2\) and \(c=(c_n/2)^2\) for some even positive integer \(c_n\) that is related to a solution of the Prouhet-Tarry-Escott problem of size \(n\), solutions of this problem are pairs of different lists \(a_1, \ldots , a_n\) and \(b_1, \ldots , b_n\) of integers such that \(\sum _{i=1}^na_i^k=\sum _{i=1}^nb_i^k\) for \(k=1, \ldots , n-1\). For such a solution the difference \(c_n=(x-b_1)\cdots (x-b_n)-(x-a_1)\cdots (x-a_n)\) is an integer that was mentioned in Theorem 1.1 above.