Algebraic sums and products of univoque bases (Q1653258)

From MaRDI portal





scientific article
Language Label Description Also known as
default for all languages
No label defined
    English
    Algebraic sums and products of univoque bases
    scientific article

      Statements

      Algebraic sums and products of univoque bases (English)
      0 references
      0 references
      0 references
      0 references
      0 references
      17 July 2018
      0 references
      Let \(x \in (0,1]\) and consider the set \({\mathcal U}(x)\) consisting of all \(\beta \in (1,2]\) such that the \(\beta\)-expansion of \(x\) is unique (the paper under review denotes \(\beta\) by \(q\)), i.e., such that there is a unique sequence \((d_i) \in \{0,1\}^{\mathbb N}\) with \(x = \sum_{i=1}^\infty d_i \beta^{-i}\). This is the set of univoque bases of \(x\). It was shown by \textit{F. Lü} et al. [Fundam. Math. 227, No. 1, 69--83 (2014; Zbl 1368.11089)] that \({\mathcal U}(x)\) is a Lebesgue null set of maximal Hausdorff dimension. In the paper under review, the authors show that for any real \(\lambda \neq 0\), the sum set \[ {\mathcal U}(x) + \lambda {\mathcal U}(x) = \{p+\lambda q : p,q \in {\mathcal U}(x)\}, \] as well as the quotient set \[ {\mathcal U}(x) \cdot {\mathcal U}(x)^\lambda = \{pq^\lambda : p,q \in {\mathcal U}(x)\}, \] contain an interval. The analogous result is also shown for the set of so-called non-matching parameters introduced by \textit{K. Dajani} and \textit{C. Kalle} [``Invariant measures, matching and the frequency of 0 for signed binary expansions'', Preprint, \url{arXiv:1703.06335}]. The proof depends on studying the associated symbolic space of sequences, using this to prove that a certain subset satisfies the property of thickness and then applying a result of \textit{S. Astels} [Trans. Am. Math. Soc. 352, No. 1, 133--170 (2000; Zbl 0967.11026)] to deduce the main result. For the special value \(x = 1\), setting \(\lambda = 1\), the main result implies that \({\mathcal U}(x) + {\mathcal U}(x)\) and \({\mathcal U}(x) \cdot {\mathcal U}(x)\) both contain an interval. However, it is shown towards the end of the paper that the sets are not themselves intervals. Nonetheless, the authors conjecture that the sets \({\mathcal U}(x) - {\mathcal U}(x)\) and \({\mathcal U}(x) \cdot {\mathcal U}(x)^{-1}\) are in fact intervals.
      0 references
      algebraic differences
      0 references
      non-integer base expansions
      0 references
      univoque bases
      0 references
      thickness
      0 references
      Cantor sets
      0 references
      non-matching parameters
      0 references

      Identifiers