On Malliavin's differentiability of BSDEs with time delayed generators driven by Brownian motions and Poisson random measures (Q988681): Difference between revisions

The authors consider the backward stochastic differential equations with time delayed generator. It is driven by a Lévy process, the components of which are given by a Brownian motion and a Poisson random measure. In these new types of equations, a generator depends in some measurable way on the past values of the solution. The approach is adopt that is based on a chaos decomposition in terms of multiple stochastic integrals with respect to the random measure to develop Malliavin's calculus for Lévy processes. In this setting, a suitable canonical space is constructed, on which a variational derivative with respect to the pure jump part of a Lévy process can be computed in a pathwise sense. The main findings are the following. it is proved that a unique solution exists, provided that the Lipschitz constant of the generator is sufficiently small, or the equation is considered on a sufficiently small time horizon. Secondly, Malliavin's differentiability of the solution of a time delayed BSDE is established, both with respect to the continuous component of the Lévy process, which coincides with the classical Malliavin derivative for Hilbert-valued random variables, as well as with respect to the pure jump part, in terms of an increment quotient operator related to Picard's difference operator.