Computational Nonlinear Stochastic Control.

The article discusses a computational methodology for solving the nonlinear stochastic optimal control based on the policy iteration algorithm and a finite dimensional approximation of the controlled Fokker-Planck-Kolomogorov/diffusion operator. It stresses the lack of stochastic equivalent to the two-point boundary value problems resulting from the Pontryagin minimum principle applied to the optimal control problem. The methodology was found to demonstrate satisfactory performance in several test cases.