Title:Progressively Enlargement of Filtrations and Control Problems for Step Processes

Abstract:In the present paper we address stochastic optimal control problems for a step process $(X,\mathbb{F})$ under a progressive enlargement of the filtration. The global information is obtained adding to the reference filtration $\mathbb{F}$ the point process $H=1_{[\tau,+\infty)}$. Here $\tau$ is a random time that can be regarded as the occurrence time of an external shock event. We study two classes of control problems, over $[0,T]$ and over the random horizon $[0,T \wedge \tau]$. We solve these control problems following a dynamical approach based on a class of BSDEs driven by the jump measure $\mu^ Z$ of the semimartingale $Z=(X,H)$, which is a step process with respect to the enlarged filtration $\mathbb G$. The BSDEs that we consider can be solved in $\mathbb{G}$ thanks to a martingale representation theorem which we also establish here. To solve the BSDEs and the control problems we need to ensure that $Z$ is quasi-left continuous in the enlarged filtration $\mathbb{G}$. Therefore, in addition to the $\mathbb{F}$-quasi left continuity of $X$, we assume some further conditions on $\tau$: the {\it avoidance} of $\mathbb{F}$-stopping times and the {\it immersion} property, or alternatively {\it Jacod's absolutely continuity} hypothesis.