Characterization of the spatial poisson process by stopping lines

From MaRDI portal





A random point field \(N\) on \(R^2\) is considered. Under some conditioning this field is proved to be a Poisson point process with the intensity measure \(\Lambda\) if and only if for any measurable increasing continuous family \((L_t)_{t\geq 0}\) of stopping lines the one-dimensional random point process \(\widetilde N\) of the form \(\widetilde N_t= N(D(L_t))\) is a doubly stochastic Poisson point process on \(R_+\) with a random parameter of intensity \(A_t=\Lambda (D(L_t))\), where \(D(L_t)= \{z\in R^2_+: z\leq L_t\}\). The stopping line \(L\) is a decreasing random curve in \(R^2\) measurable with respect to a two-dimensional filtration \((F_z)_{z\in R^2}\) where \(z\leq z'\Leftrightarrow s\leq s'\), \(t\leq t'\) \((z=(s,t))\) and \(\{z\leq L\}\in F_z\). The random parameter determines a conditional distribution of the process \(\widetilde N_t\) by the formula \[ E \bigl(e^{iu( \widetilde N_t- \widetilde N_s)} \mid\widetilde F_s\bigr)= \exp \bigl( (e^{iu}-1)(A_t-A_s) \bigr)\quad (t>s), \] where \((\widetilde F_s)_{s\geq 0}\) is the natural one-dimensional filtration of the process \(\widetilde N_t\).











This page was built for publication: Characterization of the spatial poisson process by stopping lines

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4257058)