人工神经网络耦合遗传算法优化细菌纤维素发酵培养基

Bacterial cellulose fermentation medium optimized by artificial neural network combined with genetic algorithm

细菌纤维素作为新型生物材料,已成为生物材料研究热点,但产量低限制了其进一步应用。选择合适的优化策略是实现产量提高的一个重要方法。本文在正交和均匀设计实验传统建模方法的基础上,运用人工神经网络模型结合遗传算法优化汉氏葡糖醋杆菌(Komagataeibacter hansenii HDM1-3)产细菌纤维素发酵培养基。结果表明,最佳培养基配方为葡萄糖3.98%、牛肉膏0.34%、酵母膏0.19%、磷酸氢二钠0.22%、磷酸氢二钾0.46%、乙醇2.23%。在此配方下,细菌纤维素产量最高达到2.87 g·L^(-1),与优化前培养基的产量相比提高了1.18倍。本研究优化了细菌纤维素发酵培养基参数,为细菌纤维素高产发酵及工业化推广应用提供了依据。

Bacterial cellulose as a new biomaterial has become a research hotspot,however,the low yield limited its application. Apparently,suitable optimization strategy is an important method to improve the yield. Here artificial neural networks combined with genetic algorithms were employed to optimize the fermentation medium based on traditional modeling approach of the orthogonal uniform design. The results demonstrated that the optimum formula is composed of glucose 3.98%,beef extract 0.34%,yeast extract 0.19%,disodium hydrogen phosphate 0.22%,dipotassium phosphate 0.46%,and ethanol 2.23%. With the optimal formula,the yield of bacterial cellulose was 2.87 g·L-1,1.18 times higher than that before optimization. In this research,the parameters of the fermentation medium were optimized to provide thebasis for high yield fermentation and application.