Optimal Consensus Control for Continuous-time Multi-agent Systems via Actor-Critic Neural Networks

This paper investigates the optimal consensus control problem for continuous-time multi-agent systems with switching topology by utilizing the framework of reinforcement learning. A leader-follower continuous-time high-order multi-agent system is formulated and the corresponding Hamilton-Jacobi-Bellman equation is presented. To calculate the performance index and the optimal consensus control law, a policy iteration (PI) algorithm is proposed and the convergence analysis of multi-agent systems for the algorithm is derived. Furthermore, an actor-critic neural network is applied for the PI algorithm, which does not require the knowledge of multi-agent system dynamics. A simulation example shows the effectiveness of the proposed optimal consensus control scheme.