求解非线性电阻电路的神经网络方法
Neural Network Method for Solving Nonlinear Resistive Circiuts
-
摘要: 提出的求解具有凸特性CPWL非线性电阻电路的神经网络方法,具有与常用电路分析软件相容的数据格式,且待求变量少,精度高,是一种可靠的求解方法。该方法也可推广到具有凸待性由CPWL描述的多端子非线性电阻电路的求解工作中去。Abstract: A neural network method is presented to solve two terminal nonlinear resistive circuits.The characteritics of nonlinear resitors are convex and described in canonical piece-wice linear form. The data input structure of the solftware progammed according neural net-work method is compatible with the structure of other CAA software.