Innovations two-stage dual control

摘要

A novel dual control method is proposed for the stochastic systems with unknown parameters, which converts the unsolvable dynamic programming problem into a tractable two-step ahead minimum variance control problem in a stochastic suboptimal view. Innovation variance is used to improve the learning effect, and the instant weight is introduced to reduce the influence of the future output estimation error on the system. Simulation results show the satisfactory performance of the new controller.