earticle

영어

The architecture of a class of time-varying neural networks can be determined by simply adopting that of the conventional neural networks, while the weights are allowed to vary with time. The challenge lies how to select the weights, when applying a time-varying neural network. In this paper, we use the iterative learning methodology for training time-varying neural networks, and the neural networks are proposed for modeling and identification of discrete-time time-varying nonlinear systems. Time-varying dynamical neural networks (DNNs) are presented by the architecture of conventional high-order DNNs with connection weights varying with time. Both conventional DNNs and time-varying DNNs are used to identify time-varying systems. The weights are updated by least squares integral learning algorithm with dead-zones. For time-varying case, iterative learning and its improved algorithms are used to update connection weights. The identification error is ensured to converge to the bound, which is proportional to the approximation error.