Random balls model with dependence (Q472336): Difference between revisions

The author considers a weighted random ball model in which the centers, radii and weights of balls are modeled by a Poisson random measure with intensity of the form \(f(x,r)\,dx\,dr\, G(dm)\) on \({\mathbb R}^d \times {\mathbb R}_+ \times {\mathbb R}\), and the asymptotic behavior of \(f(x,r)\) for large radii is given by \[ f(x,r) \simeq_{r\to +\infty} {g(x) \over r^{1+\beta (x)}}, \] where \(g\) and \(\beta\) are positive functions on \({\mathbb R}^d\) with \(d < \beta_1 \leq \beta (x) \leq \beta_2\). This extends the model introduced by \textit{J.-C. Breton} and \textit{C. Dombry} [Stochastic Processes Appl. 119, No. 10, 3633--3652 (2009; Zbl 1175.60049)], which corresponds to \(f(x,r) = \lambda f(r)\) and constant functions \(\beta\) and \(g\). Three different scaling regimes on the radii of balls are identified for the limiting distribution of the renormalized shot noise process, depending on the zooming factor and intensity parameters, namely, two stable integral regimes and an intermediate Poisson stochastic integral regime.