基因调控非耦联批式流加发酵切换系统的路径辨识

Pathway Identification of Switched System in an Uncoupled Fed-Batch Fermentation Based on Genetic Regulation

针对一类非耦联批式流加发酵生产1,3-丙二醇(1,3-PD)问题,综合考虑细胞内外环境、dha调节子的调控作用以及细胞内中间代谢产物3-羟基丙醛(3-HPA)对甘油脱水酶(GDHt)可能的抑制方式,建立了一个包含四条路径及四种切换模式的16维非线性切换系统模型.给出了状态变量关于参数的鲁棒性定义,并以胞外状态变量的计算值与实验值的相对误差、胞内状态变量的鲁棒性为性能指标,建立了一个非耦联批式流加发酵参数辨识模型.分析了系统状态对参数的灵敏度,基于此给出了性能指标与状态约束关于参数的梯度公式.构造了改进的并行粒子群一序列二次规划算法(PPSO-SQP).应用该算法,对辨识模型进行了数值求解,得到了数值最优的系统参数及3-HPA最可能的抑制方式.

For an uncoupled microbial fed-batch fermentation process of glycerol to 1, 3-propanediol （1, 3-PD）, a novel sixteen-dimensional nonlinear switched system model is established containing four pathways and four switched modes in consideration of comprehensive factors that both the intracellular and extracellular environments, regulation of dha regulator and the possible inhibition ways of intracellular intermediate metabolite 3-hygroxypropionaldehyde （3-HPA） to glycerol dehydratase （GDHt） are involved. The robustness of the state variables with respect to the parameters is defined. Taking the relative errors of the calculated values on the extracellular state variables and the experimental values coupling with the robustness on the intracellular state variables as the performance index, a parameter identification model for the uncoupled fed-batch fermentation is formulated. The sensitivities of the system states with respect to the parameters are analyzed, and the gradient formulas of the performance index and the state constraints with respect to the parameters are presented accordingly. An improved parallel Particle Swarm Optimization -- Sequential Quadratic Programming algorithm （PPSO-SQP） is described. The identification model is numerically solved by using the above algorithm, and the optimal system parameters and most possible inhibition way of 3-HPA are numerically determined.