基于BP人工神经网络和遗传算法的葛根总黄酮提取工艺优化研究

摘要

目的 结合Bp人工神经网络与遗传算法优化葛根总黄酮提取工艺.方法 首先利用响应面试验设计获得数据优化Bp人工神经网络模型各参数,建立相应网络模型;然后结合遗传算法通过网络进行极值寻优,获得提取工艺的最佳条件.结果 Bp人工神经网络模型的拟合度和葛根总黄酮获得率的预测结果表明,本文算法性能优于使用多元非线性回归算法性能.结论 结合遗传算法和Bp人工神经网络可优化葛根总黄酮提取工艺.%Objective:To optimize the extraction process of total flavonoids in radix puerariae with Bp artificial neural network and genetic algorithm. Methods:parameters of Bp artificial neural network model were optimized by using the data of response surface test to establish the corresponding network model,and then the network combined with the genetic algorithm was used to optimize the extreme value and obtain optimal conditions of the extraction process. Results:The fitting degree of Bp artificial neural network model and predictive results of re-ceiving rate of total flavones in radix puerariae showed that performance of the proposed algorithm was superior to that of using multiple nonlinear regression algorithms. Conclusion:Combining genetic algorithm and Bp artificial neural network to optimize the extraction process of total flavones in radix puerariae is feasible.