On a problem in simultaneous Diophantine approximation: Schmidt's conjecture (Q661926)

Throughout \(\|\cdot\|\) denotes the distance of a real number to the nearest integer. A real number is said \textit{badly approximable} if there exists a positive constant \(c(x)\) such that \(\|q(x)\|> c(x)q^{-1}\;\;\;\forall q\in \mathbb N\). Let us note \(\mathbf {Bad}\) the set of badly approximable numbers. \(\mathbf {Bad}\) is of Lebesgue measure zero. For a precise definition and details on the \textit{Hausdorff dimension} \(\dim X\) of a set \(X\) used in this article, see for instance [\textit{K. Falconer}, Fractal geometry: mathematical foundations and applications. Chichester etc.: John Wiley (1990; Zbl 0689.28003)]. A result of Jarnik states that \(\dim \mathbf {Bad}=1\). The authors use the following generalization of \(\mathbf {Bad}\) to the plane \(\mathbb R^2\): given a pair of real numbers \((i,j)\) such that \[ 0\leq i,j\leq 1\text{\;and\;}i+j=1, \tag{e2} \] a point \((x,y)\in \mathbb R^2\) is said to be \((i,j)\)-badly approximable if there exists a positive constant \(c(x,y)\) such that \(\max\{\|qx\|^{1/i},\|qy\|^{1/j}\}>c(x,y)q^{-1} \;\;\;\forall q\in\mathbb N\). The \textit{Schmidt Conjecture } (1982) is: For any \((i_1, j_1)\) and \((i_2,j_2)\) satisfying (e2) we have \(\mathbf {Bad}(i_1,j_1)\cap \mathbf {Bad}(i_2,j_2)\not=\emptyset\) [\textit{W. M. Schmidt}, Approximations diophantiennes et nombres transcendants, Colloq. Luminy/Fr. 1982, Prog. Math. 31, 271--287 (1983; Zbl 0529.10032)]. For further background, (see, \textit{H. Davenport} [Mathematika, Lond. 11, 50-58 (1964; Zbl 0122.05903)]; \textit{D. Kleinbock} and \textit{B. Weiss} [Adv. Math. 223, No. 4, 1276--1298 (2010; Zbl 1213.11148)]; \textit{S. Kristensen, R. Thorn} and \textit{S. Velani} [Adv. Math. 203, No. 1, 132--169 (2006; Zbl 1098.11039)]; [\textit{A. D. Pollington} and \textit{S. L. Velani}, Acta Math. 185, No. 2, 287--306 (2000; Zbl 0970.11026), J. Lond. Math. Soc., II. Ser. 66, No.1, 29--40 (2002; Zbl 1026.11061)]. The authors investigate the intersection of the sets \(\mathbf {Bad}(i_t,j_t)\) along fixed vertical lines in the \((x,y)\)-plane: let \(L_x\) denote the line parallel to the \(y\)-axis passing through the point \((x,0)\) and, for any real number \(0\leq i\leq 1\), define the set \[ \mathbf {Bad}(i):=\{ x\in \mathbb R : \exists\;c(x) >0\text{\;so that\;} \|qx\|>c(x)q^{-1/i}\;\;\;\forall q\in\mathbb N\}. \] The authors prove: {Theorem} Let \((i_t,j_t)\) be a countable number of pairs of real numbers satisfying ({e2}) and let \(i:=\sup\{i_t : t\in\mathbb N\}\). Suppose that \[ \liminf_{t\to\infty}\min\{i_t,j_t\}>0.\tag{e5} \] Then, for any \(\theta\in \mathbf{Bad}(i)\), we have \(\dim\Big(\bigcap_{t=1}^\infty \mathbf {Bad}(i_t,j_t)\cap L_\theta\Big)=1\). The hypothesis \(\theta\in \mathbf {Bad}(i)\) is necessary, if not the intersection of the sets \(\mathbf {Bad}(i_t,j_t)\) along the line \(L_\theta\) is empty. As a consequence, the authors prove: Let \((i_1,j_1),\dots,(i_d,j_d)\) be a finite number of pairs of real numbers satisfying (e2). Then \(\dim \Big(\bigcap_{t=1}^d \mathbf {Bad}(i_t,j_t)\Big)=2\). This theorem implies that Schmidt's conjecture is true. The authors prove also the following generalization: Let \((i_t,j_t)\) be a countable number of pairs of real numbers satisfying (e2). Suppose that (e5) is also satisfied. Then \(\dim\Big(\bigcap_{t=1}^\infty\mathbf {Bad}(i_t,j_t)\Big)=2\).