Artificial Intelligence Based Optimization of Fermentation Medium for P-Glucosidase Production from Newly Isolated Strain Tolypocladium Cylindrosporum
A Tolypocladium cylindrosporum strain was isolated for efficiently produce extracellular thermoacidophilic β-glucosidase (BGL). This objective of the present paper is to integrate two different artificial intelligence techniques namely artificial neural network(ANN) and genetic algorithm(GA) for optimizing medium composition for the production of BGL on submerged fermentations(SmF). Specifically, the ANN and GA were used for modeling non-linear process and optimizing the process. The experimental data reported in a previous study for statistical optimization were used to build the ANN model. The concentrations of the four medium components served as inputs to the ANN model and the P-glucosidase activity as the output of the model. The average error (％) and correlation coefficient for the ANN model were 1.36 and 0.998, respectively. The input parameters of ANN model were subsequently optimized using the GA. The ANN-GA model predicted a maximum |3-glucosidase activity of 2.679U/ml at the optimun medium composition. The ANN-GA model predicted gave a 22％ increase of P-glucosidase activity over the statistical optimization, which was in good agreement with the actual experiment under the optimum conditions.
Artificial intelligence(AI) Artificial neural network(ANN) Genetic algorithm(GA) fermentation medium P-glucosidase(BGL) Tolypocladium cylindrosporum
Yibo Zhang Lirong Teng Yutong Quan Hongru Tian Yuan Dong Qingfan Meng Jiahui Lu Feng Lin Xueqing Zheng
College of life science, Jilin University, Changchun 130012, China The Frist hospital of Jilin University, Changchun 130021, China
国际会议
无锡
英文
325-332
2010-09-17（万方平台首次上网日期，不代表论文的发表时间）