Functional central limit theorem for Brownian particles in domains with Robin boundary condition (Q892665): Difference between revisions

In a bounded Lipschitz domain \(D\subset \mathbb R^d\), characteristics of a diffusion process reflected on the boundary and killed near it (killing density) are specified. The initial position of the process is defined by a fixed density on \(\bar{D}\). A system of \(N\) independent particles with such dynamics is considered. The state of this particle system is a discrete measure: the empirical measure of the particles \textit{alive} on \(\bar{D}\) multiplied by \(N^{-1}\). The behavior of the corresponding measure function with the time domain \([0,\infty)\) is investigated as \(N\to \infty\), with the killing density depending on \(N\): it approximates the reflection function. Specifically, the \textit{fluctuation process} on \([0,\infty)\) is investigated: the measure function minus its expectation and multiplied by \(N^{1/2}\). The convergence result holds for any symmetric reflected diffusion, for any bounded Lipschitz domain and for any dimension \(d\geq 1\) (so the authors say, but the assumptions on the characteristics are restrictive -- non-degenerate matrix, \(W^{1,2}\) matrix and initial position density, restrictions for the killing density, specific reflection direction). The limit is a generalized Ornstein-Uhlenbeck process. The proof relies on Dirichlet form methods.