Optimization of the Extraction Process of Shenliancao Prescription Based on Box-Behnken Response Surface Design Combined with AHP-CRITIC Method and BP Neural Network-genetic Algorithm

OBJECTIVE To optimize the extraction process of Shenliancao prescription used a combination of BP neural network-genetic algorithm and AHP-CRTIC method.

METHODS This study determined the mixed weight coefficients of various indicators using the AHP-CRITIC method, with deacetyl asperulosidic acid methyl ester, scutellarin, and extract yield as indicators. Based on the Box-Behnken response surface design, the non-linear effects of water addition ratio, extraction time, and extraction frequency during the extraction process of Shenliancao prescription were reflected using BP neural network-genetic algorithm. Therefore, the optimal extraction process was determined, and the optimized results were validated through process validation.

RESULTS The BP neural network-genetic algorithm optimized the results of adding water 8 times, boiling for 1.5 h, and boiling twice, and the comprehensive score was shown to be 93.24 through validation experiments.

CONCLUSION Based on BP neural network-genetic algorithm, the extraction process of Shenliancao prescription optimized is stable and feasible. It can be effectively applied to optimize the process parameters, and also provides a new idea for the optimization of the extraction process of traditional Chinese medicine compound preparations.