Point-map-probabilities of a point process and Mecke's invariant measure equation (Q2012249)

A point-shift \(F(\phi,x)\) is an \(\mathbb{R}^d\)-valued map defined on locally finite sets \(\phi\) in \(\mathbb{R}^d\) and points \(x\) with the property that \(F(\phi,x)\in \phi\). A point-shift is compatible if \(F(\phi+t,x+t)=F(\phi,x)+t\) for all translations \(t\in\mathbb{R}^d\). Then \(F\) is uniquely determined by the associated point-map \(f(\phi)=F(\phi,0)\). The authors study point-map-probabilities (also called the \(f\)-probabilities) of a point process \(\Phi\), these probabilities can be described as the law of \(\Phi\) conditionally on the event that the origin has \(F\)-pre-images of all orders defined using iteration of the point-shift map. The authors elaborate a rigorous definition of these probabilities using tools from the theory of dynamical systems. For this, the action of the semigroup of translates of \((-f)\) on point processes is considered. The introduced point-map probabilities generalize Palm probabilities; they coincide if there is no evaporation effect, by which the authors understand the fact that the image of \(\Phi\) under iterations of \(F\) tends to the vanishing counting measure. The paper contains numerous examples of point-shifts.