In this paper, a novel event-triggered optimal control method is developed for nonlinear discrete-time systems with constrained inputs. First, a non-quadratic utility function is constructed to overcome the challenge caused by saturating actuators. Second, a novel triggering condition is designed to reduce computational burden. Difference from other triggering conditions, fewer assumptions are required to guarantee asymptotic stability. Then, the optimal cost function and control law are obtained by constructing the action-critic network. Convergence analysis of the system is provided in the consideration of the system state and neural network weight estimation errors. Finally, the effectiveness and correctness of the proposed method are verified by two numerical examples.

In this paper, a reinforcement Q-learning method based on value iteration (Ⅵ) is proposed for a class of model-free stochastic linear quadratic (SLQ) optimal tracking problem with time delay. Compared with the traditional reinforcement learning method, Q-learning method avoids the need for accurate system model. Firstly, the delay operator is introduced to construct a novel augmented system composed of the original system and the command generator. Secondly, the SLQ optimal tracking problem is transformed into a deterministic one by system transformation and the corresponding Q function of SLQ optimal tracking control is derived. Based on this, Q-learning algorithm is proposed and its convergence is proved. Finally, a simulation example shows the effectiveness of the proposed algorithm.