Optimal self-learning cooperative control for continuous-time heterogeneous multi-agent systems

In this paper, an optimal self-learning cooperative control for heterogeneous multi-agent systems by iterative adaptive dynamic programming (ADP) is developed. The main idea is to design an optimal control law by policy iteration based ADP technique which makes all the agents track a given dynamics and simultaneously makes the iterative performance index function reach the Nash equilibrium. The cooperative policy iteration algorithm for graphical differential games is developed to achieve the optimal control law for the agent of each node. Convergence properties are analyzed which make the performance index functions of heterogeneous multi-agent differential graphical games converge to the Nash equilibrium. Simulation example is given to show the effectiveness of the developed optimal self-learning control scheme.