Application of genetic algorithms to model parameter identification of a recombinant E. coli high-cell density fed-batch fermentation

Abstract

In this work, a genetic algorithm was used to estimate both yield and kinetic coefficients of an unstructured model representing a fed-batch high-cell density fermentation of E. coli. The model, based on the General State Space Dynamical Model was used to represent the three major metabolic pathways: oxidative growth on glucose, fermentative growth on glucose, and oxidative growth on acetate. The structure of the kinetic equations was derived from literature and adapted to represent experimental results. Genetic algorithms were used to minimize the normalized quadratic differences between simulated and real values of the state variables X, A and W, by manipulating both yield and kinetic coefficients.