Multi-objective Optimization Of Extraction Process Of Ligusticum Chuanxiong Hort By Fusion Of Bpartificial Neural Network And Genetic Algorithm

Objective The extraction process of traditional Chinese medicine is complex and it is difficult for traditional optimization methods to accurately reflect its multivariable nonlinear relationship.Therefore,the effective solution to the complex multi-factor and multi-level optimization problems is the key to the extraction of Chinese medicine.Ligusticum chuanxiong Hort is a commonly used traditional Chinese medicine for promoting blood circulation and removing blood stasis,but it is very difficult to further extract because of its unstable active ingredients.Therefore,it is of great theoretical and practical significance to carry out the research on the key scientific issues to optimize the extraction process of effect components of the Ligusticum chuanxiong Hort.The research results have a wide application prospect for better utilization and quality of Chinese medicine.This project takes the effect component of Ligusticum chuanxiong Hort as the research object,takes the BP artificial neural network and the genetic algorithm as the key technologies,and conducts in-depth research and analysis on the key technologies that restrict the extraction process performance of Ligusticum chuanxiong Hort effect components,in order to obtain the best extraction process of Ligusticum chuanxiong Hort.Methods Firstly,the ferulic acid and ligustilide in Ligusticum chuanxiong Hort were extracted by methanol reflux process,and the main influencing factors of methanol reflux extraction（methanol concentration,liquid-solid ratio,extraction time）were studied with uniform design.Secondly,the BP artificial neural network was introduced to establish the network model.Using ferulic acid and ligustilide as output indicators,the parameters of the BP artificial neural network model were optimized through the uniform design method.Finally,considering the multi-objective optimization characteristics of genetic algorithm,the multi-objective optimization method is used to solve the established BP artificial neural network model by using the parallel selection method.Then the design experiment is used to verify,test and evaluate the model and algorithm,thereby obtaining the best extraction process of Ligusticum chuanxiong Hort effect component.The multi-objective algorithms used in this paper are parallel selection method and NSGA-Ⅱ algorithm.Results In this paper,the data obtained from the uniform design test are fitted by linear,nonlinear and BP artificial neural networks.The results show that the multiple linear regression and nonlinear regression can not explain the relationship between the extraction parameters and the evaluation index well.The better performance of the model established by BP neural network,the correlation coefficients obtained by fitting are respectively 0.99899 and 0.99544.the parallel selection method and the non-dominated ranking genetic algorithm with elite strategy NSGA-Ⅱ were used to optimize the extraction process of Ligusticum chuanxiong Hort.The optimal extraction parameters of Ligusticum chuanxiong Hort were Methanol concentration of65%,liquid-solid ratio of 120,extraction time of 34 min,verified by the actual ferulic acid and ligustilide actual content of 2.7209 mg·g^-1,17.3229 mg·g^-1.The optimal extraction parameters of Ligusticum chuanxiong Hort by NSGA-Ⅱ algorithm were Methanol concentration of 70%,liquid-solid ratio of 120,extraction time of 35 min,and the actual contents of ferulic acid and ligustilide verified by NSGA-Ⅱ were2.7340 mg·g^-1,17.3774 mg·g^-1.Conclusion Although the process parameters obtained by the two methods are similar,the NSGA-Ⅱ algorithm shows a more stable and excellent performance in the process of multi-objective optimization by comparative analysis.Therefore,by using BP neural network combined with non-inferior genetic algorithm with elite strategy NSGA-Ⅱ to optimize the extraction process parameters of Ligusticum chuanxiong Hort,the Pareto non-inferior solution set obtained by searching is reasonable and feasible,which provides a new idea and theoretical guidance for the practical application research of traditional Chinese medicine extraction technology.