Regularity of measures induced by solutions of infinite dimensional stochastic differential equations (Q1382721)

Consider the stochastic differential equation \[ d\xi (t) = A(t, \xi (t))dw(t) + \sigma (t, \xi (t)) dt, \] where \( \xi (t) \) is a stochastic process taking values in \( R^n, W(t) \) is an \( R^n\)-valued Wiener process, \( A(\cdot, \cdot) \) is a matrix-valued function, while \(\sigma (\cdot,\cdot)\) is a vector-valued function. It is well known that under some conditions involving the smoothness of \( A\) and \(\sigma\) with respect to the second argument, the density of the transition probability of the solution satisfies the Kolmogorov forward equation. In particular, it is smooth enough. The main aim of the paper is to extend these results to the infinite-dimensional case. One of the principal difficulties here is related with the fact that in infinite-dimensional spaces there is no good analog of the Lebesgue measure, and so one cannot speak of the density of a transition probability. All the more, the smoothness of the density loses its sense. It has been noticed by \textit{V. I. Averbukh, V. I. Smolyanov} and \textit{S. V. Fomin} [Trans. Moscow Math. Soc. 24(1971), 140-184 (1974); translation from Tr. Moskov. Mat. Obshch. 24, 133-174 (1971; Zbl 0234.28005)] that in the finite-dimensional case the smoothness of the density of a measure \(\mu\) can be expressed in terms of \(\mu\) only. Namely it is equivalent to the smoothness of the function \(f_A(h) = \mu (A+h) \) for every Borel set \(A\). And the three authors came up with the idea to consider this smoothness of the density of \(\mu\) in the infinite-dimensional case as an analog the smoothness of the density of \(\mu\). In the paper under review the author develops the theory of differentiable measures to obtain regularity of measures induced by solutions of stochastic integral equations. The method used by the author is similar to that used in the finite-dimensional case by \textit{K. Bichteler} and \textit{D. Fonken} [in: Stochastic processes. Prog. Probab. Stat. 5, 97-110 (1983; Zbl 0538.60058)].