Multi-Level Programming Models And Solution Methods For Computing Objective Function Of Glycerol Metabolism

1, 3-propanediol is an important chemical raw materials, and it has important applications for excellent performance of polymer synthesis monomers. So microbial fermentation production of 1, 3-propanediol is widespread concerned of scholars. However, the fermentation process is a relatively difficult nonlinear optimization system. We should establish and develop adaptive optimization model and method of this metabolic process to optimize the production of 1, 3-propanedionl process to provide theoretical guidance.Glycerol metabolism production of 1, 3-propanediol as a background, the paper researches on computing objective function of glycerol.The main content and results obtained of thesis are as follows:1. In order to investigate maximize biomass, maximize ATP, minimize flux squares, minimize ATP and maximize flux sum, the paper made five metabolic objective function as underlying objective function and made the squares of flux values and experimental flux values as upper objective function. Considering the constraints of flux balance analysis, this paper established five different objective functions of bi-level optimization model. In response to these bi-level programming, this paper uses duality principle of linear and nonlinear K-T conditions and nonlinear mixed integer programming to convert into a single-level programming problem, and analyzes the results. In contrast, the objective function is maximum ATP with satisfactory results. But having a great deviation between the optimal values of the flux values through the experimental values, it is necessary to find a more suitable glycerol metabolism in the objective function.2. In response to these five objective functions would not be a good description of metabolic, this paper established a multi-level programming model of glycerol metabolism. In order to solve the multi-level programming model, the paper put forward an effective method for optimization. First, the multi-level programming model is changed into an bi-level programming model. Secondly, using theoretical analysis make it transform into the corresponding 11 sub-problem. Finally, the iterative algorithm is designed to solve this multi-level programming problem. The results showed that, compared with the existing literature, this paper not only the calculated values, but also establishes a precise metabolic objective function.