基于GA-LM-BP神经网络的锂离子电池预测研究

Research of Li-Ion Battery Prediction Based on GA-LM-BP Neural Network

摘要:

针对BP神经网络训练学习速度慢、容易陷入局部极小值的缺陷，利用LM算法融合高斯-牛顿法和梯度下降法优点的快速性，充分利用遗传算法全局随机搜索强的优势，构建了三层5-6-1型的GA-LM-BP神经网络结构，优化BP神经网络的初始权值和阀值，减少了BP神经网络陷入局部极小值的几率。通过对锂离子电池数据进行了实验，结果表明了该方法预测的有效性。

Abstract:

The training learning speed of BP neural network is slow, and is easy to fall into local minimum value, the paper uses LM algorithm fusion Gauss - Newton method and the quickness of gradient descent method advantages. It is global random search strong advantage by making full use of the Genetic Algorithm, to build the three layers of the type 5-6-1 - LM - structure of BP neural network. The GA optimizes the BP neural network's initial weights and thresholds, reduces the risk of BP neural network into a local minimum value. The results show the validity of the proposed method.