Markov properties for point processes on the plane (Q923492)

In 1948, P. Lévy defined the sharp Markov property in the following most natural way: We say that the process \(\{X_ z\), \(z\in {\mathbb{R}}^ 2\}\) has the sharp Markov property with respect to a set A if the \(\sigma\)-fields \(\sigma \{X_ z\); \(z\in A\}\) and \(\sigma \{X_ z\); \(z\in A^ c\}\) are conditionally independent given \(\sigma \{X_ z\); \(z\in \partial A\}\). Lévy concluded by asking the question: Do there exist non-degenerate two-parameter processes with the sharp Markov property with respect to a sufficiently large class of sets? A weaker definition, called the germ field Markov property was introduced in 1963 by McKean. In this definition, the \(\sigma\)-field \(\sigma \{X_ z\); \(z\in \partial A\}\) is replaced by the larger germ \(\sigma\)-field \(\cap_{\epsilon >0}\sigma \{X_ z\); \(d(z,\partial A)<\epsilon \}.\) Most of the papers on Markov random fields deal with the Gaussian case; the literature on the Markov property for point processes is very limited and started with an unpublished paper of Carnal 1980, in which it is proved that the Poisson process satisfies the sharp Markov property relative to all bounded open relatively convex sets. \textit{F. Russo} [Sémin. Probabilités XVIII, 1982/83, Proc., Lect. Notes Math. 1059, 353-378 (1984; Zbl 0537.60025)] extended the Carnal result in the following sense: Every two-parameter point process with independent increments satisfies the sharp Markov property relative to any finite union of rectangles whose sides are parallel to the axes. In this paper it is proved that for a wide class of point processes on the plane, and for a wide class of compact sets A the germ \(\sigma\)-field \(\cap_{\epsilon >0}\sigma \{X_ z\); \(d(z,A)<\epsilon \}\) is equal to the \(\sigma\)-field \(\sigma \{X_ z\); \(z\in A)\). Using this result, it is shown that any point process with independent increments and absolutely continuous intensity which is strictly simple has the sharp Markov property relative to any open connected and bounded set with a piecewise decreasing or increasing boundary. An additional section of this paper is devoted to an investigation of the strong Markov property of point processes, that is the sharp Markov property relative to measurable and adapted random sets.