Mean-square random dynamical systems (Q439108)

Two-parameter semigroups operating on the space \(L^2(\Omega,P;\mathbb R^d)\) of square-integrable \(\mathbb R^d\)-valued random variables are introduced in order to investigate random or stochastic differential equations with non-local coefficients. Assuming exponential contraction of differences of trajectories in the course of time, uniform in the initial time and locally uniform in~\(L^2\), existence of a limiting point process in~\(L^2\) is shown. This can be interpreted as a `one point attractor' for the two-parameter semigroup, both pullback and forward (see [\textit{A. N.\ Carvalho}, \textit{J. A.\ Langa} and \textit{J. C.\ Robinson}, Attractors for infinite-dimensional non-autonomous dynamical systems. Berlin: Springer (2013; Zbl 1263.37002)]). Two one-dimensional examples of stochastic differential equations are discussed. For general situations problems with this approach arise from the difficulty of characterising compactness in \(L^2\).