Cuckoo Search Algorithm for Parameter Identification of Fermentation Process Model

Parameter identification of non-linear dynamic processes, among them fermentation ones, is rather difficult and non-trivial task to be solved. Failure of conventional optimization methods to provide a satisfactory solution provokes the idea some stochastic algorithms to be tested. As such, the promising metaheuristic algorithm Cuckoo search (CS) has been adapted and applied for a first time to a parameter identification of S. cerevisiae fed-batch fermentation process model. Aiming to improve the model accuracy and the algorithm convergence time, several pre-tests adjustments of CS have been done according to the specific optimization problem. Obtained results confirm the effectiveness and efficacy of the applied CS algorithm. In addition, a comparison between CS and simple genetic algorithm, proved as successful in parameter identification of fermentation process model, has been done. Algorithms advantages and disadvantages have been outlined and the more reliable one have been distinguished.