An approach to solve nonlinear H∞ control problem based on neural networks

In this paper, a neural network approach to solve nonlinear H_∞ control problem is proposed. An approximation solution of the Hamilton-Jacobi inequality can be obtained after neural network online learning along dynamic trajectories of closed-loop system sufficiently, then the state feedback H_∞ controller is explicitly realized by the neural networks. In order to ensuring the stability of system during learning procedure of network, a learning algorithm combined with linear H_∞ control is given