基于MOGA和SVM的发酵过程建模

Modeling of Fermentation Process Based on MOGA and SVM

摘要: 针对微生物发酵过程，使用多目标遗传算法（multi-objective genetical gorithm，MOGA）确定最优参数．MOGA和SVM回归相结合形成一种新的建模方法，该方法利用现场生产数据建立了青霉素效价预估模型．仿真结果表明此方法具有很强的拟合和泛化能力．MOGA方法的有效性也得到了验证，它也能够自动选择最优参数．

Abstract: Multi-objective genetic algorithm(MOGA) is used to determine the optimal parameters of the microbial fermentation process.A new modeling method that combines MOGA with support vector machine(SVM) regression is presented,with which the penicillin titer pre-estimation model is developed with data collected from real plant.Simulation results show that this method possesses strong capability of fitting and generalization.The effectiveness of MOGA is validated,and it can select parameters automatically.