神经网络和遗传算法优化细菌纤维素培养参数

Optimization of culture parameters for bacterial cellulose by artificial neural network and genetic algorithm

以果葡糖浆和玉米浆为主要原料,依次通过Plackett-Burman设计、均匀设计、人工神经网络和遗传算法,纤维素产生菌G-29静态培养下的最优条件是:果葡糖浆40g/L,玉米浆23.88g/L,(NH4)2SO43g/L,Na2HPO4·12H2O 4.66g/L,柠檬酸1g/L,醋酸0.5g/L,MgSO4·7H2O 0.2g/L,FeSO4·7H2O 0.4g/L,CaCl20.015g/L,ZnSO4·7H2O0.004g/L,乙醇2%(v/v),接种量10%(v/v),种龄2d,初始pH 5.82,培养时间5.94d,培养温度30℃.

Using high fructose syrup （HFS）and corn steep liquor （CSL）,the fermentation medium of bacterial cellulose （BC）was optimized by Plackett-Burman design （PB）,uniform design （UD）,artificial neural network （ANN）and genetic algorithm （GA）successively.The optimal culture parameters were：HFS 40 g/L,CSL 23.88 g/L,（NH4）2SO43 g/L,Na2HPO4·12H2O 4.66 g/L,citric acid 1 g/L,acetic acid 0.5 g/L,MgSO4·7H2O 0.2 g/L,FeSO4·7H2O 0.4 g/L,CaCl2 0.015 g/L,ZnSO4·7H2O 0. 004 g/L,ethanol 2% （v/v）,inoculum size 10% （v/v）,inoculum age 2 d,initial pH 5.82,culture peri-od 5.94 d at 30℃.