Hausdorff dimension of the image of iterated additive processes (Q990323): Difference between revisions

Let \(\{X^j_{T_j}\}\), \(t_j\in \mathbb{R}^1_+\quad 1\leq j\leq N\) be independent additive processes in \(\mathbb{R}^d_+\) and \(e^{\Psi_{t_j}(\xi )}\), \(\xi\in \mathbb{R}^d\) be the corresponding characteristic function. An iterated additive process \(X_t\) in \(\mathbb{R}^d\) is the sum \[ X_t=X^1_{t_1}+\ldots + X^N_{t_N}, \quad (t_1, t_2, \ldots , t_N)\in \mathbb{R}^N_+. \] Let \(X(G)=\{X_t:teG\}\) for a Borel set \(G\in \mathcal{B}(\mathbb{R}^N_+)\). Let \(P(G)\) denote the collection of probability measures on \(G\). Let \(S^{\alpha}_t=S^1_{t_1}+S^2_{t_2}+\ldots + S^d_{t_d}\) be the sum of \(\alpha\)-stable Lévy processes in \(\mathbb{R}^d\) with the Lévy exponent \(|\xi |^{\alpha}, \quad 0<\alpha < 1\). For any probability measure \(\mu\) in \(\mathbb{R^N_+}\) and \(\xi\in \mathbb{R}^d\) denote \[ \begin{aligned} & Q_{\mu}(\xi )\\ &=\int_{\mathbb{R}^N_+}\int_{\mathbb{R}^N_+}e^{-\sum^N_{j=1}\text{sgn}(s_j-t_j)[\Psi^j_{sj}(\text{s g n}(s_j-t_j)\xi)-\Psi^j_{tj}(\text{sgn} (s_j-t_j)\xi)]}\mu (ds)\mu (dt). \end{aligned} \] The main result is Th. 1.1. on the intersection probability \[ P\Big\{ X(G)\bigcap S^{1-\beta /d}((0, \infty )^d)\neq \varnothing \Big\} >0\Longleftrightarrow \int_{\mathbb{R}^d}|\xi |^{\beta -d} Q_\mu (\xi ) d\xi <\infty \] for some \(\mu \in P(G)\), for all \(\beta \in (0, d)\) and \(S^{1-B/d}\) independent of \(X\). The expression for the Hausdorff dimension \(\dim_HX(G)\) is consequence of Th.11. The volume of the paper is 21 page. The list of references contains 2 positions.