Dynamic Optimization Of Mixed Integer And Its Application In Glycerol Fermentation

This dissertation studies the nonlinear differential dynamical systems and their optimal control problems with the bio-dissimilation of glycerol to 1,3-propanediol by klebsiella pneumoniae in the background.The contents include the nonlinear differential dynamical systems of the microbial continuous and batch cultures、system identification and optimal control.The main results,obtained in this dissertation,may be summarized as follows:1.This paper focuses on modeling and optimal control of the continuous conversion process of glycerol to 1,3-propanediol in which the dilution rate of the medium is constantly changed with time.A nonlinear three phase hybrid system is presented to describe this bioconversion process.For the purpose of improving the mean productivity of 1,3-propanediol by optimizing the dynamic dilution rate,we propose a discrete-valued optimal control problem,which is transformed into a continuous optimal control problem via a time transformation method that eventually becomes a large-scale parameter optimization problem by discretization.A multi-strategy competition artificial bee colony is constructed to solve this problem.Besides,a quasi-rounding strategy is also developed aiming to generate the optimal switching sequence based on the obtained numerical solution.Computer simulations show the rationality of the three-phrase hybrid system and the feasibility of the dynamic dilution rate.And numerical results indicate that the developed numerical algorithm is efficient and valid for this complex applied optimization problem.2.The fermentation of glycerol by klebsiella pneumoniae is a complex bioprocess.Especially,the regulatory mechanisms can change constantly over time due to internal and external perturbations.In this paper,a generalized nonlinear hybrid dynamical system is proposed to describe this batch fermentation process under the assumption that 1,3-propanediol passes the cell membrane by passive diffusion coupled with facilitated transport and the transport mechanism of glycerol is unknown.To infer the most reasonable dynamic transport mechanism of glycerol,a system identification model consisting of both discrete and continuous variables is proposed,in which the cost function minimizes the defined biological robustness.The partial outer convexification and the time-scale transformation are used to convert this problem to a large-scale parameter optimization problem.In consideration of both the difficulty of finding analytical solutions and the highly complex nature of this problem,we develop a modified artificial bee colony algorithm with heuristics of tabu search to solve this model.Finally,we explore the appropriateness of the optimal dynamic transport mechanism of glycerol as well as the effectiveness of the algorithm via numerical simulations.