Policy iteration for optimal control of weakly coupled nonlinear systems with completely unknown dynamics
Date of Original Version
In this paper, an online learning algorithm based on policy iteration is established to solve the optimal control problem for weakly coupled nonlinear continuous-time systems. Using the weak coupling theory, the original problem is transformed into three reduced-order optimal control problems. To obtain the optimal control laws without system dynamics, we construct an online data-based integral policy iteration algorithm which is used to solve the decoupled optimal control problems. The actor-critic technique based on neural networks and the least squares method are used to implement the model-free learning algorithm. A simulation example is given to verify the applicability of the developed algorithm.
Proceedings of the American Control Conference
Li, Chao, Ding Wang, Derong Liu, and Haibo He. "Policy iteration for optimal control of weakly coupled nonlinear systems with completely unknown dynamics." Proceedings of the American Control Conference 2016-July, (2016): 5722-5727. doi:10.1109/ACC.2016.7526566.