A recurrence criterion for Markov processes of Ornstein-Uhlenbeck type (Q1123486): Difference between revisions

A Markov process of Ornstein-Uhlenbeck type was introduced by \textit{K. Sato} and \textit{M. Yamazato} [see e.g. Stochastic Processes Appl. 17, 73- 100 (1984; Zbl 0533.60021)] as a Markov process on \(R^ d\) generated by a Lévy process generator plus a drift term \(- \sum^{d}_{j=1}\sum^{d}_{k=1}Q_{jk}x_ k \partial /\partial x_ j\) with the matrix \(Q=(Q_{jk})\) having eigenvalues with positive real parts. A criterion for positive recurrence of processes of this type was given by Sato and Yamazato. This paper gives a criterion for null recurrence and transience by an integral condition involving the Lévy measure in the case of one dimension. Multi-dimensional cases are also discussed.