An improved differential evolution algorithm for enhancing biochemical pathways simulation and production

This paper presents an Improved Differential Evolution (IDE) algorithm to improve the kinetic parameter estimation in simulating the glycolysis pathway and the threonine biosynthesis pathway. Experimentally derived time series kinetic data are noisy and possess many unknown parameters. These characteristics of kinetic data cause lengthy computational time to compute the optimum value of the kinetic parameters. To solve this problem, this study had been conducted to develop a hybrid method that combined the Differential Evolution algorithm (DE) and the Kalman Filter (KF) to produce IDE. Results have shown that lesser computation time (6% and 18.5% faster) and more robust to noisy data with significant reduced error rates (93% and 79% reduced error rates) compared with the Genetic Algorithm (GA) and DE, respectively, in glycolysis and threonine biosynthesis pathway simulations. IDE is reliable as it demonstrated consistent standard deviation values which were close to mean values. We foresee the applicability of IDE into other metabolic pathway simulations.