Robust optimal control for a consumption-investment problem (Q2482684)

The author considers the problem of maximizing utility from both terminal wealth and intertemporal consumption under model uncertainty. The latter is assumed to be of the type, that the prior possible model is characterized by a probability measure belonging to the class \[ \mathcal{Q} = \left\{ \mathbf{Q} \sim \mathbf{P} \left| \frac{d \mathbf{Q}}{d\mathbf{P}} = \mathcal{E} \left( \int_0 \eta_{1t} dW_t^1 + \int_0 \eta_{2t} dW_t^2 \right)_T , (\eta_1,\eta_2) \in \mathcal{C} \right. \right\} \] and the investors optimal control problem is given by \[ \max \inf_{\mathbf{Q} \in \mathcal{Q}} \mathbf{E}_\mathbf{Q} \left[ \int_0^T \gamma e^{-\lambda t} U(c_t) dt + U(X_T^{x,c,\pi})\right] \] where \(U\) denotes a utility function of HARA class with risk aversion parameter \(\alpha >0\). The author derives a PDE and proves that it characterizes the value function of the optimal control problem.