Invariant probabilities for systems in a random environment - with applications to the Brusselator (Q799605)

Analytical conditions for the existence and uniqueness of an invariant probability for a Markov process (x(t,\(\omega)\),y(t,\(\omega)\)) that arises as a solution to the system \(x'(t)=f(x(t),y(t,\omega))\) are provided by Khasminskij and Papanicolaou. The author also has addressed this problem jointly with L. Arnold by comparing the above stochastic system with a deterministic control system \(x'(t)=f(x(t),u(t)).\) The purpose of the author in this article is to give further (sufficient) conditions for the existence-uniqueness of invariant distribution for (x,y) through the existence-uniqueness of invariant control sets for (x,u). He then applies his result to 2-dimensional stochastically perturbed Brusselator chemical reaction equations.