Application of model-predictive control based on artificial neural networks to optimize the fed-batch process for riboflavin production

The fed-batch process for commercial production of riboflavin (vitamin B-2) was optimized on-line using model-predictive control based on artificial n eural networks (ANNs). The information required for process models was extr acted from both historical data and heuristic rules. After each cultivation the process model was readapted off-line to include the most recent proces s data. The control signal (feed rate), however, was optimized on-line at e ach sampling interval. An optimizer simulated variations in the control sig nal and assessed the forecasted model outputs and the amount according to a n objective function. The optimum feed profile for increasing the product y ield (Y-B2/S) of riboflavin at the time of harvesting was adjusted continuo usly and applied to the process. In contrast to the control by set-point pr ofiles, the novel ANN-control is able to react on-line to variations in the process and also to incorporate the new process information continuously. As a result, both the total amount of riboflavin produced and the product y ield increased systematically by more than 10% and the reproducibility of s even subsequently optimized batches was enhanced. (C) 2000 Elsevier Science B.V. All rights reserved.