论文部分内容阅读
通过对经典BP网络算法的改进,引入迭代步长优化的自适应规则,有效地避免了迭代解的振荡问题,提高了收敛速度。工程实例应用表明,基于模糊神经网络模型反演参数的方法具有精度高、收敛快等优点。
By improving the classical BP network algorithm, the adaptive rule of iterative step optimization is introduced, which effectively avoids the oscillation problem of iterative solution and improves the convergence speed. The application of engineering examples shows that the method based on fuzzy neural network inversion parameters has the advantages of high precision and fast convergence.