Materials, Methods & Technologies, Volume 5, 2011

Abstract: The paper presents the problems of real parameter optimization of fermentation process mathematical models. General and global optimisation methods are classified here and some short notes on the comparison of different methods are shown. A typical test function, as well as two mathematical models of E. coli MC4110 fed-batch fermentation processes are considered as case studies. Considering used test function the results show that the all optimization methods, except Genetic algorithms, get into the local minimum. Considering fermentation process models identification simulation results reveal that accurate and consistent results can be obtained only using genetic algorithms. The best result is obtained using the Multipopulation genetic algorithm.