An explicit bound for aliquot cycles of repdigits (Q2911033)

Let \(s(n)\) be the sum of the divisors of \(n\) smaller than \(n\) and let \(s_k(n)\) be the \(k\) fold iteration of \(n\). Then \(n\) is said to be in aliquot cycle if \(s_k(n)=n\) for some \(k\) and the numbers \(n,s(n),s_2(n),\dots,s_{k-1}(n)\) are the members of the aliquot cycles. The reviewer and te Riele showed that given an integer base \(g\geq 2\) which is even, there are only finitely many aliquot cycles consisting of positive integers which are all rep digits in base \(g\); namely, numbers of the form \(d(g^m-1)/(g-1)\) for some digit \(d\in \{1,\dots,g-1\}\) in base \(g\) and \(m\geq 1\). In the paper under review the author makes this result explicit by showing that the largest member of such a cycle is at most \(T_{2g+2}(g)\), where \(T(x)=g^{14g x}\), and, as before, \(T_{\ell}(x)\) is the \(\ell\)th fold iteration of \(T\) evaluated in \(x\). The proof follows closely the arguments of the reviewer with te Riele. Such arguments are based on finiteness results from the theory of diophantine equations. The author makes such steps entirely explicit thanks to prior results by Baker, Nagell, Rotkiewicz and Luca and Pollack.