Application of artificial bee colony algorithm for model parameter identification

Abstract

In this chapter, the Artificial bee colony (ABC) algorithm, based on the foraging behaviour of honey bees, is introduced for a numerical optimization problem. The ABC algorithm is one of the efficient population-based biological-inspired algorithms. To demonstrate the usefulness of the presented approach, the ABC algorithm is applied to parameter identification of an E. coli MC4110 fed-batch cultivation process model. The mathematical model of E. coli MC4110 cultivation process is considered as a system of three ordinary differential equations, describing the two main process variables, namely biomass and substrate dynamics, as well as the volume variation. This case study has not been solved previously in the literature by application of ABC algorithm. To obtain a better performance of the ABC algorithm, i.e. high accuracy of the solution within reasonable time, the influence of the algorithm parameters has been investigated. Eight ABC algorithms are applied to parameter identification of the E. coli cultivation process model. The results are compared, based on obtained estimates of model parameters, objective function value, computation time and some statistical measures. As a result, two algorithms are chosen—ABC1 and ABC8, respectively, with 60 × 500 number and 20 × 400 (population × maximum cycle number), such as algorithms with the best performance. Further, the best ABC algorithms are compared with four population-based biological-inspired algorithms, namely Genetic algorithm, Ant colony optimization, Firefly algorithm and Cuckoo search algorithm. The results from literature of metaheuristics applied for the considered here parameter identification problem are used. The results clearly show that the ABC algorithm outperforms the biological-inspired algorithms under consideration, taking into account the overall search ability and computational efficiency.