摘要
Kriging代理模型常常作为工程计算领域的一种优化方法,通过选取动态点能有效提高Kriging代理模型的精确度,这种方法在优化多目标问题上具有高精度的特点,本实验设置了酒曲微生物混菌10组不同比例(V_细∶V_酵∶V_霉=1∶0∶0、0∶0∶1、0∶1∶0、1∶1∶1、4∶4∶2、4∶2∶4、2∶4∶4、2∶2∶6、2∶6∶2、6∶2∶2)进行实验,混合液态发酵7 d后,以5种不同的不饱和脂肪酸(肉豆蔻油酸、棕榈油酸、油酸、亚油酸、α-亚麻酸)甲酯化后的含量为基础点进行Kriging方法模型建立。通过优化模型的建立选取了3组混合比例(V_细∶V_酵∶V_霉=1.6∶2.7∶5.7、1.7∶3.9∶4.4、1.9∶1.9∶6.2)进行验证实验,当预测值的含量与验证实验真实值的相对偏差小于10%时,说明此方法可行。通过验证实验,选取V_细∶V_酵∶V_霉=1.9∶1.9∶6.2作为最终的优化结果。
Kriging model, as a commonly-used optimization method in engineering calculation field, its accuracy could be improved effectively by selecting dynamic points. This method is featured by high precision in the optimization of multiple targets. In the experiment, ten groups of microbial strains of different mixed ratio (Vbacteria:Vyeast:Vmold= 1 : 0: 0, 0: 0: 1, 0: 1 : 0, 1 : 1 : 1, 4: 4: 2, 4: 2: 4, 2: 4: 4, 2: 2: 6, 2: 6: : 2, 6: 2: 2) were set up, 7 d after mixed liquid fermentation, the Kriging models had been established with the content of five kinds of unsaturated fatty acids after esterification (fourteenolefinic acid methyl, palm methyl oleate, methyl oleate, linoleic acid methyl ester, alpha linolenic acid) as the basic points. The optimized Kriging models were then verified in three groups (mixed ratio of Vbacteria:Vyeast:Vmold=1.6:2.7: 5.7, 1.7:3.9:4.4, 1.9: 1.9: 6.2). As the deviation of predicted values and real values was less than 10 %, the model was feasible. In conclusion, the mixed ratio Vbacteria:Vyeast:Vmold= 1.9 :1.9 : 6.2 was the best choice in practice.
出处
《酿酒科技》
2016年第9期23-27,共5页
Liquor-Making Science & Technology
基金
董香型白酒工艺创新技术合作项目
关键词
KRIGING模型
混合菌株
液态发酵
不饱和脂肪酸
比例优化
Kriging model
mixed microbial strains
liquid-state fermentation
unsaturated fatty acids
ratio optimization