基于模糊对向网络的非线性动态系统辨识新算法

New algorithm for identification of fuzzy CP network based nonlinear dynamic systems

以模糊对向网络为辨识模型 ,利用它所具有的模糊逻辑系统和神经网络两者的优点 ,在模糊C -均值聚类方法的基础上引入新的聚类目标函数 ,并证明了它的优化条件 .将模糊聚类和最小二乘法相结合 ,提出一种模型参数学习的新算法 .在直流电弧燃弧时间建模研究中的应用结果表明提出的算法是有效的 .

Presents the model used for identification of nonlinear system by using FPC network with the characteristics of both fuzzy logic systems and neural networks, which presents a new fuzzy clustering object function based on FCM, with optimized condition proved, and a new learning algorithm of model parameters developed by combining fuzzy clustering with LS and concludes from the application results in research of modeling for DC arc duration that the proposed learning algorithm is effective.