Iterated sums of fifth powers of digits (Q1011052): Difference between revisions

Let \(e\geq 1\) and \(b\geq 2\) be integers. An \textit{\(e\)-power \(b\)-happy number} is a positive integer \(n\) having the property that some iterate of the map that sends \(n\) to the sum of the \(e\) powers of its base \(b\) digits is \(1\). When \(e=2\) and \(b=10\) these are referred to as simply \textit{happy} numbers. In [Rocky Mt. J. Math. 30, No. 2, 565--570 (2000; Zbl 1052.11008)], \textit{E. El-Sedy} and \textit{S. Siksek} proved that there are arbitrarily large strings of consecutive happy numbers. In the paper under review, the authors extend this result to \(e=5\) and various bases \(b\). First the authors show that any \(5\)-power \(b\)-happy number must necessarily be congruent to \(1\) modulo \(d=\gcd(30,b-1)\). In particular, in base \(10\) any \(5\)-power happy number must be congruent to \(1\) modulo \(3\). A second important result in the paper is that there are infinitely many bases \(b\geq 2\) (including all numbers \(b\in \{2,3,\dots,10\}\)) for which there are arbitrarily long strings of \(5\)-power \(b\)-happy numbers which are consecutive terms of the arithmetic progression \(1\) modulo \(d\). The proofs are elementary but nice.