On the arithmetic of cycles for the Collatz-Hasse (`Syracuse') conjectures (Q1099201)

For \(n\) an odd natural number, let \(f_{\epsilon}(n)=2^{- k}(3n+\epsilon)\), where \(k\) is the highest power of 2 dividing \(3n+\epsilon\). The Syracuse problem is to show for all \(n\) some iterate of \(f_ 1(n)\) equals 1. The author also considers the settling of the iterations of \(f_{-1}(n)\) into one of the three cycles (i) 1,1,...; (ii) 5,7,5,... or (iii) 17,25,37,55,41,61,91,17,.... His principal theorem establishes (a) a 1-1 correspondence between an integer \(n\) in a cycle and the cycle length vector of powers \(k\) defined by the iterates in the cycle starting at \(n\); and (b) the equivalency of the existence of such a vector and three conditions related to a cyclic group of permutations acting on a direct sum of natural numbers. The interesting corollary is: if the above conjectures are true then the equation \(2^{\ell}-3^ r=\epsilon\) has only the positive integral solution (2,1) if \(\epsilon =1\), and the positive integral solutions (1,1) or (3,2) if \(\epsilon =-1\). (Compare to a result in \textit{R. Tijdeman}'s ''On the Fermat-Catalan equation,'' J.-ber. Dtsch. Math. Ver. 87, 1-18 (1985; Zbl 0577.10016)]. One of the conditions of the theorem translates into an explicit computational criterion for determining cycle vectors. A final theorem gives an upper bound on a natural number to be in a cycle given its cycle vector.