基于神经网络和遗传算法的醋酸发酵培养基优化

The Optimization of Fermentation Medium of Acetic Acid Bacteria Based on Neural Network and Genetic Algorithm

为优化醋酸菌As1.41发酵培养基,采用正交试验确定最佳碳、氮源种类,并利用神经网络和遗传算法寻求最佳培养基浓度。试验结果:最佳碳、氮源为葡萄糖与酵母膏;最优组合是:葡萄糖2.7%,酵母膏3.1%,乙醇6.6%,NaCl0.41g/100mL。在接种量4%,温度30℃,转速120r/min下摇床发酵5d,做对照试验,结果表明:利用神经网络和遗传算法优化的培养基产酸量比优化前提高了31%,比正交试验优化结果提高了16%,说明神经网络与遗传算法在培养基优化中具有显著的优越性。

The fermentation medium of acetic acid bacteria is optimized in this paper.The optimum carbon and nitrogen sources is decided by orthogonal test,and the optimum concentration is obtained using neural network and genetic algorithm.The results of the experiment show that the optimum carbon and nitrogen sources are glucose and yeast extract;the optimal medium contents are glucose 2.7% and yeast extract 3.1% and ethanol 10.6% and NaCl 0.11 g/100 mL.A contrast experiment is designed between the two optimal fermentation mediums obtained by GA and the orthogonal test and the original medium under the condition of inoculums 4% and temperature 30 ℃ and 120 rpm.The result indicates that the acidity of the optimal fermentation medium obtained by GA is increased by 31% than the original medium and 16% than the medium obtained by the orthogonal test.It can be show that using neural network and genetic algorithm have a remarkable superiority in medium optimization.