Mixed Neural Network and Its Application to the Control of Nonlinear Systems

A parallel neural network—mixed neural network formed by nonline ar and linear neural networks is put forward in accordance of the characteristic s of a nonlinear dynamic system model.The new neural network solves the problems of complexity and uncertainty caused by nonlinear function approximating linear function.The respective approximation of the linear and the nonlinear parts for model have been realized.The simulation examples demonstrate that the applicati on of the mixed neural network to the control of this nonlinear dynamic system i s feasible.