Equidistribution of expanding measures with local maximal dimension and Diophantine approximation (Q766211)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Equidistribution of expanding measures with local maximal dimension and Diophantine approximation
scientific article

    Statements

    Equidistribution of expanding measures with local maximal dimension and Diophantine approximation (English)
    0 references
    0 references
    23 March 2012
    0 references
    Let \(m\), \(n\) be positive integers and denote by \(M_{m,n}= \mathbb{R}^{mn}\) the space of \(m\times n\) matrices with real entries. Dirichlet's Theorem (DT) on simultaneous Diophantine approximations says that given \(Y\in M_{m,n}\) and \(N\geq 1\), there exist \(q= (q_1,\dots, q_m)\in \mathbb{Z}^m\setminus\{0\}\subset M_{1,m}\) and \(p= (p_1,\dots, p_n)\in \mathbb{Z}^n\subset M_{1,n}\) with \(\| qY+ p\|\leq{1\over N^m}\) and \(\| q\|\leq N^n\). Here \(\|.\|\) stands for sup norm on \(\mathbb{R}^k\). Given \(Y\) as above and a positive number \(\sigma< 1\), we say that DT can be \(\sigma\)-improved for \(Y\) and write \(Y\in DI_\sigma(m, n)\) or \(Y\in DI_\sigma\) when the dimensions are clear from the context, if for every \(N\) large enough one can find \(q= (q_1,\dots, q_m)\in\mathbb{Z}^m\setminus\{0\}\) and \(p= (p_1,\dots, p_n)\in\mathbb{Z}^n\) with \(\| qY+ p\|\leq{\sigma\over N^m}\) and \(\| q\|\leq\sigma N^n\). We say that DT can be improved for \(Y\) if \(Y\in DI_\sigma\) for some \(0<\sigma< 1\). The following theorem of Davenport and Schmidt says that DT can not be improved for most \(Y\): For any \(m,n\in\mathbb{N}\) and any positive number \(\sigma< 1\), the set \(Y\in DI_\sigma(m, n)\) has Lebesgue measure zero. In case \(m=1\), Kleinbock and Weiss showed that for a large class of measures DT can not be \(\sigma\)-improved for almost every element if \(\sigma<\sigma_0\) for some positive \(\sigma_0\) depending on the measure. After that Shah improved the result by removing the upper bound \(\sigma_0\) for a special kind of measures concentrated on analytic curves. The aim of the paper under review is to generalize the theorem of Davenport and Schmidt in a direction in some sense opposite to the theorem of Shah. Instead of a smooth one-dimensional submanifold, measures supported on a full Hausdorff dimension subset of \( M_{m,n}\) are considered and shown that for almost every point DT cannot be improved. Using entropy and dynamics on homogeneous spaces of Lie groups it is shown that for a certain class of fractals \(K\subset[0,1]^{mn}\subset \mathrm M_{m,n}(\mathbb R)\) of local maximal dimension DT cannot be improved almost everywhere.
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    Dirichlet's theorem
    0 references
    Lebesgue measure
    0 references
    entropy
    0 references
    0 references
    0 references