On supports of evolution systems of measures for converging in law non-homogenous Markov processes (Q2179269)

The work investigates connections of the stochastic evolution of a Markov-Feller nonhomogenous process with the asymptotic behavior of deterministic set-valued semigroup induced by the considered process. The author gives an explicit form of supports of strongly mixing evolution system of measures naturally connected with nonhomogenous Markov processes induced by time-dependent stochastic partial differential equations. Sections 2 and 3 present set-valued processes of lower semi-continuous multifunctions. Section 4 introduces the notion of two deterministic families of sets associated with such processes, namely an evolution semi-attractor and the core of an evolution semi-attractor. Their topological properties are studied. The main result of the paper (Theorem 5.6 in Section 5) shows that the core of an evolution semiattractor of a two-parameter semigroup of Markov multifunctions appears to be supports of a strongly mixing evolution system of measures of two-parameter semigroup of Markov-Feller linear operators.