Validation and Analyses of General DTRNN with Different Values of Selection Parameter for Solving Discrete-Form Time-Variant Complex Division

In recent years, many engineering problems have gradually evolved into discrete-form time-variant problems, and the researches show the superior performance for solving above discrete-form time-variant problems in discrete-time environment using recurrent neural network (RNN) method. In this paper, a discrete-time RNN (DTRNN) model is established by using the general single-parameter square-pattern discretization (SPD) formula for solving discrete-form time-variant complex division. Specifically, first of all, according to the traditional Euclid division, the continuous-form time- variant complex division is presented, and the corresponding discrete-form time-variant complex division is established and converted. Then, the presented discretization method and the corresponding DTRNN model are developed. Finally, numerical experiments show that the effective results of general DTRNN model with different values of selection parameter.