Leader-Following Constrained Distributed Adaptive Dynamic Programming Design for Multiagent Systems

This paper gives an adaptive dynamic programming (ADP)-based distributed adaptive control scheme to solve the cooperative control problem with input constraints. To compensate the effects of the constrained-input, a proper nonquadratic functional is selected to encode the saturation nonlinearity into the optimization formulation. By constructing the single network to estimate the solution of coupled nonlinear Hamilton-Jacobi-Bellman (HJB) equation, distributed cooperative optimal control law can be obtained, which can make the nonzero-sum (NZS) games reach the Nash equilibrium. In addition, the updating law of each NN is designed and implemented simultaneously. Finally, the local consensus error and the estimation error of the NN weight are proved to be boundedness. A numerical simulation is given to verify the effectiveness of the developed method.