Finite-Time Optimal Control of Polynomial Systems Using Successive Suboptimal Approximations