Broyden-Method Aided Discrete ZNN Solving the Systems of Time-Varying Nonlinear Equations

Recently, a special class of recurrent neural network (termed Zhang neural network, ZNN) has been generalized for solving systems of time-varying nonlinear equations (STVNE), and a resultant continuous (or say, continuous-time) ZNN model has been proposed and analyzed. To generalize the idea for digital computers and numerical algorithms, this paper discretizes the continuous STVNE-solving ZNN using Euler difference and improves the discrete (or say, discrete time) ZNN models by employing Broyden method. Results of various numerical experiments are presented to verify the effectiveness of the proposed discrete ZNN models, especially the Broyden-method aided ones.