$3x+1$ inverse orbit generating functions almost always have natural boundaries

TL;DR

This paper studies the generating functions related to the $3x+1$ problem, showing they almost always have the unit circle as a natural boundary, which relates to the conjecture about their rationality.

Contribution

It provides sufficient conditions under which the generating functions for $3x+1$ orbits have the unit circle as a natural boundary, linking this to the $3x+1$ conjecture.

Findings

Most generating functions have the unit circle as a natural boundary.

For certain initial values, the functions may be rational, relating to the conjecture.

The results connect analytic properties of generating functions to the $3x+1$ problem.

Abstract

The $3x+k$ function $T_{k}(n)$ sends $n$ to $(3n+k)/2$ resp. $n/2,$ according as $n$ is odd, resp. even, where $k \equiv \pm 1~(\bmod \, 6)$. The map $T_k(\cdot)$ sends integers to integers, and for $m \ge 1$ let $n \rightarrow m$ mean that $m$ is in the forward orbit of $n$ under iteration of $T_k(\cdot).$ We consider the generating functions $f_{k,m}(z) = \sum_{n>0, n \rightarrow m} z^{n},$ which are holomorphic in the unit disk. We give sufficient conditions on $(k,m)$ for the functions $f_{k, m}(z)$ have the unit circle $\{|z|=1\}$ as a natural boundary to analytic continuation. For the $3x+1$ function these conditions hold for all $m \ge 1$ to show that $f_{1,m}(z)$ has the unit circle as a natural boundary except possibly for $m= 1, 2, 4$ and $8$. The $3x+1$ Conjecture is equivalent to the assertion that $f_{1, m}(z)$ is a rational function of $z$ for the remaining values $m=1,2, 4, 8$.