摘要
采用响应面法和人工神经网络耦合遗传算法,对影响固态发酵产纳豆激酶的工艺参数进行了优化,并对这两种方法的优化效果进行评价。结果表明:在纳豆激酶固态发酵工艺参数优化中,采用人工神经网络耦合遗传算法较响应面法具有更好的数据拟合能力和预测准确度,纳豆激酶固态发酵工艺最佳参数:接种量4%,初始含水量55%,大豆装量90 g/250 mL,发酵温度36.09℃,蔗糖添加量1.5%,MgSO4.7H2O添加量0.21%,CaCl2添加量0.27%,发酵时间24 h。在该条件下发酵产物的最大酶活可达7 631.28±219.54 U/g,较单因素试验的最高水平提高了29.02%。
Response surface methodology and artificial neural network coupling genetic algorithm were used for optimizing solid-state fermentation process parameters of nattokinase,and the optimization effects of two methods were estimated.The results showed that the artificial neural network combined with genetic algorithm possessed higher fitness and prediction accuracy in the solid-state fermentation process.The best conditions were as following: inoculum size 4%,initial water amount 55%,soybean content 90 g/250 mL,fermentation temperature 36.09 ℃,sucrose concentration 1.5%,MgSO4·7H2O concentration 0.21%,CaCl2 concentration 0.27%,fermentation time 24 h.Under these conditions,the maximum nattokinase activity was up to 7631.28±219.54 U/g,which was 29.02% higher than that in single factor test.
出处
《食品与发酵工业》
CAS
CSCD
北大核心
2013年第1期134-141,共8页
Food and Fermentation Industries
基金
河北省科技攻关计划(11275508D)
关键词
纳豆激酶
固态发酵
响应面法
人工神经网络
遗传算法
nattokinase
solid-state fermentation
response surface methodology
artificial neural network
genetic algorithm