Online optimization of a fed-batch bioreactor

The mathematical optimization of growth conditions of biochemical systems like bacterial and fungal systems in a bioreactor has been the focus of research for past many years. We propose a simple but working model of the process. The dynamic optimization of the process was carried out with adjustment of model parameters at each sampling time. An essential component of online optimization problem solver is an optimization subroutine that finds global optimum fast enough to be effective for a dynamic process. Also, the information of the most sensitive model parameter and an algorithm that deals with process nonlinearities while taking care of process model mismatch are required. A method called heuristic random optimizer was used for online optimization of the process, together with the fuzzy logic based algorithm for updating the model parameters. This scheme was implemented on a fed-batch reactor and a hybridoma cell line 4W4 was studied. A significant improvement in the product yields was observed using the above-mentioned approach.