The divergence of Banach space valued random variables on Wiener space (Q1779995)

Let \((W,H,\mu)\) be an abstract Wiener space, i.e. \(W\) is a separable Banach space, \(H\) a Hilbert space densely embedded in \(W\), and \(\mu\) denotes a Gaussian probability measure on \(W\) such that every continuous linear form \(l\) on \(W\) is a centered Gaussian random variable with variance \(| l| _H^2\). The spaces are conventionally assumed to be embedded as follows: \(W^*\subseteq H\subseteq W\subseteq W^{**}\). The authors first present a definition of divergence \(\delta\) of a \(W^{**}\)-valued random variable in such a way that it coincides with the standard definition for \(H\)-valued random variables, and several propositions related to the divergence are proven. In particular, it is shown that Sobolev spaces are contained in the domain of \(\delta\), each element \(v\) in the domain of \(\delta\) can be uniquely decomposed as a sum of a divergence free element \(v^0\) and an exact element \(v_e\) (the so called Shigekawa's decomposition), and a complete characterization of divergence free \(W^{**}\)-valued random variables is given. Next, the divergence of random variables with values in operators from \(W^*\) to a Banach space \(Y\) is introduced, and its domain is characterized in case \(Y\) is a second dual to some Banach space. The rest of the paper is devoted to applications of the results to analysis of Wiener space valued vector fields generating flows on \(W\).