基于HGA-BP算法的调速系统辨识研究

针对BP网络学习算法的不足,提出了一种基于递阶遗传算法（HGA）的BP网络建模方法。文中对递阶遗传算法在优化神经网络过程中的一些参数,如适应度函数、编码方案和交叉变异概率等提出具体的设计方法。该算法可以同时对神经网络进行拓扑结构优化和参数求解。最后采用该方法,实现了具有复杂非线性特征的某调速系统的精确建模。实验结果及其与改进BP算法建模方法的比较,验证了该算法的有效性。

Identification of Speed Governing System Based on HGA-BP

For the shortage of BP neural network learning algorithm,a method for modeling BP network was proposed based on the hierarchical genetic algorithm.This paper presented the specific design method of the parameters,such as the fitness function,coding,crossover and mutation probability in the neural network optimization using hierarchical genetic algorithm.This method can optimize topological structure and parameters of the neural network at the same time.Finally,the technique was employed to settle the modeling problem of the speed governing system of a certain weapon with complex non-linear characteristics.The experimental results and comparisons with modeling methods of BP algorithm indicated that the proposed approach can be more effectively to construct precise models.