Geometric series with randomly increasing exponents

From MaRDI portal
Publication:2447692





The author proves some results on absolute continuity and singularity of the distribution \(\mu_{\beta}\) of a power series with randomly increasing exponents \(X_{\beta}\). Let \(X\) be a random variable taking values in \(\mathbb N\), and set \(p_i=\operatorname{P}(X=i)\) for \(i\in {\mathbb N}\). We assume that \(X\) has finite expectation value \(\operatorname{E}(X)\) and entropy \(H(X)\). Then the author proved for all \(\beta \in (0,1)\) with \(\beta < e^{-H(X)/\operatorname{E}(X)}\), that \(\nu_{\beta}\) is singular with respect to the Lebesgue measure; and that for \(\lambda\)-almost all \(\beta \in (e^{-H(X)/\operatorname{E}(X)},0.668)\), \(\mu_{\beta}\) is absolute continuous with respect to the Lebesgue measure. The author also proved that \(\dim_H \mu_{\beta}\leq \frac{H(X)}{-\operatorname{E}(X)\log \beta}\), where \(\dim_H \mu_{\beta}\) is the Hausdorff dimension of \(\mu_{\beta}\). Last, for \(q\in (1,2]\) and well defined \(\beta_q \in (0,1)\), the author shows for almost all \(\beta \in (\beta_q,0.668)\), that \(\mu_{\beta}\) is absolutely continuous with density in \(L^q\).











This page was built for publication: Geometric series with randomly increasing exponents

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2447692)