This paper investigates nonlinear stochastic optimal control problems under non-Gaussian white noise, which is represented by so-called Lévy process. We show that the path integral approach combined with policy iteration gives a numerical solution to the Hamilton–Jacobi–Bellman equation.
