Anticipating Stochastic Differential Systems with Memory

This article establishes existence and uniqueness of solutions to two classes of stochastic systems with finite memory subject to anticipating initial conditions which are sufficiently smooth in the Malliavin sense. The two classes are semilinear stochastic functional differential equations (sfdes) and fully nonlinear sfdes with a sublinear drift term. For the semilinear case, we use Malliavin calculus techniques, existence of the stochastic semiflow and an infinite-dimensional substitution theorem. For the fully nonlinear case, we employ an anticipating version of the Itô-Ventzell formula due to Ocone and Pardoux ([O-P]). In both cases, the use of Malliavin calculus techniques is necessitated by the infinite-dimensionality of the initial condition.