Hausdorff dimension of Gauss-Cantor sets and two applications to classical Lagrange and Markov spectra (Q2089665)

From MaRDI portal





scientific article
Language Label Description Also known as
default for all languages
No label defined
    English
    Hausdorff dimension of Gauss-Cantor sets and two applications to classical Lagrange and Markov spectra
    scientific article

      Statements

      Hausdorff dimension of Gauss-Cantor sets and two applications to classical Lagrange and Markov spectra (English)
      0 references
      0 references
      0 references
      0 references
      0 references
      24 October 2022
      0 references
      Let \[ L := \Big\{\limsup_{\substack{p,q \to \infty \\ p \in \mathbb{Z},q \in \mathbb{N}}} \frac{1}{\lvert q^2\alpha - pg\rvert} < \infty: \alpha \in \mathbb{R}\Big\} \] and \[ M := \Big\{\sup_{\substack{(p,q) \in \mathbb{Z}^2\\ (p,q) \neq (0,0)}} \frac{1}{\lvert ap^2 + bpq + cq^2\rvert} < \infty: ax^2 + bxy + cy^2 \text{ real indefinite, } b^2 - 4ac = 1\Big\} \] denote the classical Lagrange respectively Markov spectrum and let \(\dim_{H}(X)\) denote the Hausdorff dimension of \(X\). As it was shown in [\textit{C. Moreira}, Ann. Math. (2) 188, No. 1, 145--170 (2018; Zbl 1404.11095)], for any \(t > 0\), \(\dim_{H}((-\infty,t]\cap M) = \dim_{H}((-\infty,t]\cap L)\) and \(f(t) := \dim_{H}((-\infty,t]\cap M)\) is continuous and non-decreasing. In the first main statement (Theorem 2.1), the authors consider \(t_1 := \inf\{t \in \mathbb{R}: f(t) = 1\}\) and prove that \(t_1 = 3.3334384\ldots.\) This confirms heuristic bounds obtained by \textit{R. T. Bumby} [J. Reine Angew. Math. 331, 192--206 (1982; Zbl 0468.10032)] and improves the currently best rigorous upper and lower bounds by \textit{M. Hall jun.} [Acta Arith. 18, 387--399 (1971; Zbl 0224.10023)] and \textit{C. Moreira} [Ann. Math. (2) 188, No. 1, 145--170 (2018; Zbl 1404.11095)] given by \(\sqrt{10} = 3.162277\ldots < t_1 < \sqrt{12} = 3.464101\ldots\). By a similar approach, the values \(\inf\{t \in \mathbb{R}: f(t) = \frac{n}{k}\}, k = 1,2,\ldots,n-1\) are approximated, with the approximate plot of \(f\) being shown in the article (Figure 1). The second main statement (Theorem 3.1) treats the Hausdorff dimension of \(M\setminus L\). The authors show that \(0.537152 < \dim_{H}(M \setminus L) < 0.796445\) which improves the currently best bounds \(0.5312 < \dim_{H}(M \setminus L) < 0.8823\) due to \textit{C. Matheus} and \textit{C. G. Moreira} [Comment. Math. Helv. 95, No. 3, 593--633 (2020; Zbl 1465.11165)] and \textit{M. Pollicott} and \textit{P. Vytnova} [``Hausdorff dimension estimates applied to Lagrange and Markov spectra, Zaremba theory, and limit sets of Fuchsian groups'', Trans. Am. Math. Soc., Ser. B 9, 1102--1159 (2022; Zbl 07638529)]. Both proofs are computer-assisted and could be refined when using additional computational resources.
      0 references
      0 references
      Markov and Lagrange spectra
      0 references
      Hausdorff dimension
      0 references
      continued fraction algorithm
      0 references
      subshifts of finite type
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references