BP 算法在磁化曲线拟合中的应用

BP Algorithm for Fitting the Excitation Characteristic Curve

在优化学习率ＢＰ算法的基础上，根据磁化曲线的特点，进行了一些改进与简化．提出了一种简化的快速算法，并编制了训练程序．结果表明，该算法可较大幅度地提高拟合精度，而且节省较多的计算时间．

The optimum learning rate BP(back propagation) algorithm is improved and simplified in accordance with the excitation characteristic curve. The scheme is used to design a training program whereby both the fitting accuracy and calculating time are enhanced.