Martingale solutions and invariant measures for stochastic evolution equations in Banach spaces (Q1613661): Difference between revisions

The factorization procedure, proposed by \textit{G. Da~Prato, S. Kwapień} and \textit{J. Zabczyk} [Stochastics 23, 1-23 (1987; Zbl 0634.60053)] as a method of establishing regularity of trajectories of mild solutions to stochastic evolution equations in Hilbert spaces, proved itself a useful tool in other problems as well. The second author and \textit{B. Gołdys} [Stochastics Stochastics Rep. 46, No. 1-2, 41-51 (1994; Zbl 0824.60052)] used it to find martingale solutions of a semilinear stochastic partial differential equation \[ dX = (AX + F(t,X)) dt+ G(t,X) dW \tag{1} \] in a Hilbert space, where \(A\) is a generator of a compact \(C_0\)-semigroup and \(F\), \(G\) are continuous functions of a linear growth. \textit{G. Da Prato}, the second author and \textit{J. Zabczyk} [Stochastic Anal. Appl. 10, No. 4, 387-408 (1992; Zbl 0758.60057)] employed the factorization method to give sufficient conditions, in the autonomous case, for the existence of invariant measures for (1), again under the linear growth assumption on \(F\) and \(G\). The aim of the paper under review is to generalize these two results, first, by replacing the linear growth of \(F\) by a suitable dissipativity hypothesis, and second, by allowing more general state spaces than the Hilbert space (namely, type 2 UMD Banach spaces are considered). The very basic scheme of proofs remains the same as in the papers cited above, however, the details are quite different and rather delicate analysis is required.