基于梯度提升回归树的发酵过程软测量研究

Research on Soft Measurement of Fermentation Process Based on Gradient Boost Regression Tree

生物发酵过程的优化控制需要很多变量的参与,其中有许多无法实现实时在线测量。对此,论文结合群体智能算法和机器学习技术,提出了一种基于梯度提升回归树的发酵过程软测量方法,同时利用果蝇算法对回归树的关键参数进行寻优,以及偏移补偿技术对模型输出值进行校正。以Pensim仿真平台所得的青霉素发酵数据检验论文方法的效果,结果表明,该方法的预测精度在5.42%以内,明显优于BPNN、SVR、AdaBoost和RF方法,对发酵过程的关键变量预测有指导意义。

Optimal control of the biological fermentation process requires the participation of many variables,many of which are unable to achieve real-time online measurements.Therefore,combining with swarm intelligence algorithm and machine learning technology,this paper proposes a soft-measurement method for fermentation process based on gradient-lifting regression tree.At the same time,the key parameters of the regression tree are optimized by fruit fly optimization algorithm,and the offset compensa⁃tion technique is used to correct the output value of the model.The model performance of penicillin fermentation data obtained by Pensim simulation platform is verified.The results show that the accuracy of the algorithm is within 5.42%,which is better than BPNN,SVR,AdaBoost and RF algorithms,with guiding significance for the prediction of key variables in fermentation process.