基于神经网络和遗传算法的黄芪发酵培养基优化

为了提高乳杆菌FGM9发酵黄芪产物中的粗多糖得率,采用均匀设计法对发酵培养基各组分配比进行初步探讨,运用人工神经网络法建立了黄芪发酵液中多糖得率与培养基各组分浓度之间的预测模型,通过遗传算法对该模型进行优化和评估,得出发酵培养基各组分间的最优配比。结果显示,组分A含量为42.10 g/L、蛋白胨7.45 g/L、酵母膏9.80 g/L、KH2PO4-K2HPO42.18g/L、组分E 6.25 g/L的条件下,发酵液粗多糖的得率最高可达25.6%,比优化前提高了6.85%。

Optimization of Radix astragalus fermentation medium based on neural network and genetic algorithms

To improve the crude polysaccharide（CP）yield in the produets of Radix astragalus fermentation by Lactobacillu FCM9,the homogeneous design was used for the trial of the formula of medium components ,neural network was applied for the pre diet modeling of the relationship between the CP yield and the concentration of medium components,the genctic alg orithm was used for the global optimization of the fermentation model.The optimized culture medium was composed of constiuent A 42.10g/L,peptone 7.45g/L,yeast extract 9.80g/L,KH2PO4-K2HPO4 1.80g/L and constituent E 6.25g/L.CPyield wasimproved to 25.6% and increased by 6.85% compared to the original medium.