非支配排序遗传算法多目标优化补阳还五汤总苷提取的实验研究

Study of glycosides extraction of multi-objective optimization BYHWD by non-dominated sorting genetic algorithm

目的:多目标优化补阳还五汤总苷的提取条件。方法:采用BP人工神经网络建立补阳还五汤中黄芪甲苷、芍药苷和苦杏仁苷提取工艺的多目标优化模型,依据非支配排序遗传算法(NSGA-Ⅱ)对提取条件进行多目标优化。结果:得出了多目标优化模型的Pareto最优解集,纳入决策者偏好确定最佳提取条件为:溶剂量10.94倍,提取1.26h,提取3次,提取温度82.06℃。结论:NSGA-Ⅱ方法与BP人工神经网络结合可对补阳还五汤总苷提取条件实现有效的多目标寻优,本方法可为医药学研究领域中的多目标优化问题提供基于人工智能算法的解决方案。

Objective： To establish a multi-objective optimization BYHWD extraction of total glycosides.Methods： The BP artificial network multi-objective optimization model was adopted to build up the extraction of glycosides of BYHWD by the artificial neural network.An intelligent method for solving multi-objective optimization problems（MOP） was used in extraction solutions based on nondominated sorting genetic algorithm（NSGA-II）.Results： Obtained Pareto multi-objective optimization model for optimal solution set,and the optimum conditions was as follows： liquid-solid ratio was 10.94,extraction time was 1.26h,temperature was 82.06℃ and extracting three times.Conclusion： BP artificial neural network and NSGA-II are very suitable for BYHWD extraction of total glucosides of multi-objective optimization.This method can be a reference to the MOP in the field of medicine study based on artificial intelligence algorithm.