Solving Bisymmetric Solution of a Class of Matrix Equations Based on Linear Saturated System Model Neural Network

In order to solve the complicated process and low efficiency accuracy of solving a class matrix equations, this paper introduces linear saturated system model neural network architecture bisymmetric solution equations. Firstly, equations is constructed determine key problems Secondly, structure characteristic parameters in solution. Then, solved by using backpropagation topology. Finally, normalization realized objective function solution, realized. verify effect method paper, three indexes (accuracy, correction accuracy, time) are designed experiment. The experimental results show that proposed can effectively reduce time, improve has high practicability.