VIII: 06, 27-36, LNM 381 (1974)
DINGES, Hermann
Stopping sequences (
Markov processes,
Potential theory)
Given a discrete time Markov process $(X_n)$ with transition kernel $P$, a stopping sequence with initial distribution $\mu$ is a family $(\mu_n)$ of measures such that $\mu\ge\mu_0$ and $\mu_{k-1}P\ge\mu_k$. The stopping sequence associated with a stopping time $T$ is the sequence of distributions of $X_{T}, k< T<\infty$ under the law $P_\mu$. Every stopping sequence arises in this way from some randomized stopping time $T$, and the distribution of $X_T, T<\infty$ is independent of $T$ and called the final distribution. Then several constructions of stopping sequences are described, including Rost's ``filling scheme'', and several operations on stopping sequences, aiming at the construction of ``short'' stopping times in the Skorohod imbedding problem, without assuming transience of the process
Comment: This is a development of the research of H.~Rost on the ``filling scheme'', for which see
523,
524,
612. This article contains announcements of further results
Keywords: Discrete time Markov processes,
Skorohod imbedding,
Filling schemeNature: Original
Retrieve article from Numdam
