Optimal operation of bioprocesses is extremely important because these processes utilize expensive raw materials, require large capital-intensive plants and yield products that are low in concentration. Due to the inherent complexity of biological systems, reliable mathematical models needed for optimization and control of the fermentations are seldom available. Any change occurring in the environment would alter the model and thus the optimal feeding strategy calculated off-line may not be optimal for the exact process. Thus, in spite of the availability of elegant theories on dynamic optimization and control, there has been very limited success in implementing these theories experimentally in fermentations. This study will develop a concept of on-line optimization for batch and fed-batch bioprocesses that takes into account the continually changing environmental conditions and unmodeled intracellular events using a mathematical process description with time-varying parameters. A scheme will be developed which estimates the model parameters on-line and computes and updates the optimal feeding policy while the fermentation is in progress. The work will address specifically the direct application of the proposed optimization concept to an industrially significant process, namely cephalosporin fermentation, and will try to validate the results through experimentation.
