Abstract
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. Mathematical optimization requires the availability of a process model. Greater the accuracy of the process model, more reliable the optimization results are. Lack of process models with high accuracy has always been a bottleneck for such an optimization problem. More so, there are so many different cell types being used in the biotech industry today that modeling each of them rigorously on individual basis is a daunting task. This problem has been addressed by proposing a simple but working model of the process. The dynamic optimization of the process is carried out with adjustment of model parameters at each sampling time. An essential component of online optimization problem solver will be an optimization subroutine that finds global optimum fast enough to be effective for a dynamic process. Also the information about most sensitive model parameters and an algorithm that could take care of process nonlinearities while taking care of process model mismatch is required. A method called Heuristic Random Optimizer (HRO) (Li and Rhinehart 1998) has been used for on-line optimization of the process. Fuzzy logic based algorithm has been used for updating the model parameters. This scheme has been implemented on a fed-batch reactor and a hybridoma cell line 4W4 has been studied upon. A Significant improvement in the product yields has been observed using the above-mentioned mechanism.
| Original language | English |
|---|---|
| Pages (from-to) | 2986-2990 |
| Number of pages | 5 |
| Journal | Proceedings of the American Control Conference |
| Volume | 5 |
| State | Published - 1999 |
| Event | Proceedings of the 1999 American Control Conference (99ACC) - San Diego, CA, USA Duration: Jun 2 1999 → Jun 4 1999 |