Discrete-time Zhang neural network and numerical algorithm for time-varying linear equations solving

Abstract

Since 2001, a special class of recurrent neural network (RNN), termed Zhang neural network (ZNN), has been proposed, generalized and exploited for online solution of time-varying problems by following Zhang et al's design method. In this paper, for possible digital hardware realization, discrete-time ZNN models are proposed and investigated for time-varying linear equations solving. Such discrete-time ZNN models make use of time-derivative information of the time-varying coefficients. For comparison purposes, a discrete-time gradient-based neural network (GNN) model is also presented to solve the same problem. Simulative and numerical results illustrate the efficacy and superiority of these discrete-time ZNN models used for time-varying linear equations solving, as compared to the GNN model.