On finitely many base $q$ expansions
Simon Baker, George Bender
TL;DR
This work provides the first explicit families of intervals in $(1,2)$ for which bases $q$ admit a nonempty set of points with exactly $m$ base-$q$ expansions and, moreover, positive Hausdorff dimension for all $m\ge3$. The authors develop a quantitative framework built from Cantor-set thickness, interleaving, and the Falconer–Yavicoli intersection theorem to locate these intervals around the $k$-Bonacci bases $q_k$, and they obtain explicit lower bounds on the dimension of the corresponding $\mathcal{U}_q^{(m)}$. A refined result for $\mathcal{B}_3$ uses a constructive set $A_q$ and Newhouse’s thickness-interleaving criterion to produce concrete neighbourhoods near certain $q_k$ and near $q_9$, with explicit Hausdorff-dimension guarantees. The methods bridge dynamical base-$q$ expansions, geometric measure theory, and fractal-intersection techniques, offering a robust path to quantify the size and location of multi-expansion bases in $(1,2)$.
Abstract
Given some integer $m \geq 3$, we find the first explicit collection of countably many intervals in $(1,2)$ such that for any $q$ in one of these intervals, the set of points with exactly $m$ base $q$ expansions is nonempty and moreover has positive Hausdorff dimension. Our method relies on an application of a theorem proved by Falconer and Yavicoli, which guarantees that the intersection of a family of compact subsets of $\mathbb{R}^d$ has positive Hausdorff dimension under certain conditions.
