微生物连续发酵建模及胞内动力学参数辨识

Modeling of microbial continuous fermentations and identifying of intracellular kinetic parameters

针对甘油连续发酵生产1,3-丙二醇过程,建立了涉及胞内物质及跨膜运输方式的动力学系统,并讨论了该系统的一些性质.以计算值与实验数据的平均相对误差作为目标函数,给出了估计胞内动力学参数的参数辨识模型,并证明了该辨识模型最优参数的存在性.最后构造了改进粒子群算法求解辨识模型中的最优参数.数值结果表明:平均相对误差减小了20.635%,且该算法比其他算法具有更好的全局收敛性.

A dynamical system concerning intracellular substances and transport patterns was proposed to describe glycerol bioconversion to 1,3-propanediol in continuous fermentation process.Some properties of the system were subsequently discussed.Taking the error between the computational values and experimental data as the cost function,a parameter identification model to identify the intracellular kinetic parameters was established.The existence of the optimal parameters in the parameter identification model was also proved.Finally,an improved particle swarm optimization was constructed to solve the above parameter identification model.Numerical results show that the mean relative error is cut down by 20.635%,and the proposed algorithm has higher global convergence compared with other optimization algorithms.