Poisson cluster measures: Quasi-invariance, integration by parts and equilibrium stochastic dynamics (Q1004369)

The paper deals with a characterization of Poisson cluster point processes in \(X=\mathbb{R}^d\). The Poisson cluster measure is introduced with the aid of a certain Poisson measure on an auxiliary configuration space over the disjoint union \(\mathfrak{X}=\bigsqcup_n X^n\) with intensity measure obtained as a convolution of the backgropund intensity of the cluster centers with the probability distribution of a generic cluster. It is shown that the Poisson cluster measure is quasi-invariant with respect to compactly supported diffeomorphisms of \(X\) and an integration-by-parts formula is established. An irreducible Dirichlet form associated with the Poisson cluster measure is constructed and the corresponding equilibrium stochastic dynamics is derived. In contrast to the usual setting [cf. \textit{K. Matthes}, et al. Infinitely divisible point processes. (Chichester) etc.: John Wiley \& Sons. (1978; Zbl 0383.60001)] or [\textit{D. J. Daley}, et al. An introduction to the theory of point processes. Vol. I. Probability and Its Applications. (New York), NY: Springer. (2003; Zbl 1026.60061)]) the configuration space is not restricted to locally finite configurations. The approach and the results should be compared also with [\textit{A. Liemant}, et al. Equilibrium distributions of branching processes. Mathematical Research, 42. (Berlin) (GDR): Akademie-Verlag. (1988; Zbl 0671.60076)].