Stochastic and robust control of nonlinear economic systems (Q1330535)

A robust optimal policy formulation to decision making under uncertainty for nonlinear economic systems is considered (robust policy anticipates the possibility of a disturbance). The method used is an extension of the mean-variance approach usual for linear models to nonlinear systems and is in the context of parametrized simple feedback decision rules. The parameters are determined to minimize a convex combination of the original objective function and a sensitivity expression. The expectations of nonlinear systems can be calculated by evaluating the bias from certainty equivalent controls using Monte Carlo simulations. The sensitivity expression describing the robustness of the control can be chosen either as the variance of the original objective function or of the individual endogenous variables. Two numerical examples of the robust policy method are given.