Identification of parameters in chemical kinetics using a hybrid algorithm of artificial bee colony algorithm and simplex

Abstract

Parameter identification is a key step in establishing kinetic models. Aimed at the above problem, it can be transformed into an optimization problem by constructing objective function that minimizes simulation errors. Two examples of chemical kinetics are analyzed. Among them, the second case is a more complicated nonlinear problem, so using ABC individually can not obtain better results. Then a hybrid algorithm of ABC and simplex is proposed. The method firstly uses ABC to carry out global search so as to obtain better initial point. Secondly, simplex is employed to process local search based on the above initial point. Compared with that of modified genetic algorithm, hierarchical differential evolution, adaptive differential evolution, the results show that the new hybrid algorithm can obtain better optimization precision by combining global searching ability of ABC with strong local searching ability of simplex. So it is an effective optimization method. © 2011 Springer-Verlag.