On layered stable processes (Q880486): Difference between revisions

Let \(\mu\) be an infinitely divisible probability measure without Gaussian component. It is called a layered stable (\(LS_{\alpha,\beta}(\sigma,q,\eta)\)) if its Lévy measure on \(\mathbb{R}^n\backslash \{0\}\) is \[ \nu(B)=\int_{S^{d-1}} \sigma(d\xi)\int_0^\infty 1_B(r\xi)q(r,\xi)\, dr, \] where \(\sigma\) is a finite positive measure on \(S^{d-1}\), and \(q:(0,\infty)\times S^{d-1}\to (0,\infty)\) is a locally integrable function, such that \[ q(r,\xi)\sim c_1(\xi) r^{-\alpha-1}, \text{ as }r\to 0, \quad \text{and}\quad q(r,\xi)\sim c_2(\xi) r^{-\beta-1}, \text{ as } r\to\infty, \] for \(\sigma\)-a.e. \(\xi\), and \(c_1(\xi)\) and \(c_2(\xi)\) are positive integrable on \(S^{d-1}\) functions. The Lévy process with Lévy measure \(\nu\) is called layered stable \(LS_{\alpha,\beta}(\sigma, q,\eta)\), where \(\eta\) is a drift. The indexes \(\alpha\) and \(\beta\) are called the inner and outer indexes of \(\mu\). The authors show that the outer index is responsible for the existence of moments, while the variational properties are determined by the inner index. Further, after suitable re-scaling near zero, the \(LS_{\alpha,\beta}(\sigma, q,0)\) process weakly converges to a \(\alpha\)-stable process with Lévy measure depending on \(\sigma_1\), while after suitable re-scaling near infinity, the process weakly converges to a \(\beta\)-stable process with Lévy measure depending on \(\sigma_2\) if \(\beta<2\), or in case \(\beta>2\) it converges to a centered Brownian motion with some covariance matrix. In relation to the short-time behavior, the mutual absolute continuity of a layered stable process and of its short-time limiting stable process is investigated. Finally, the shot-noise series representations of layered processes are discussed.