Data-Driven Optimal Control for Continuous- Time Linear Nonzero-Sum Games

In this paper, the data-driven optimal control problem is studied for continuous-time linear nonzero-sum games. Two kinds of reinforcement learning algorithms, i.e., reinforcement learning algorithm with data-storage based least-square method and reinforcement learning algorithm with filter based least-square method, are presented to obtain the Nash equilibrium solution. The properties of the presented rein-forcement learning algorithms are analyzed. Simulation results show the efficiency of the presented reinforcement learning algorithms.