Robust multi-objective optimal control of uncertain (bio)chemical processes

Dynamic optimization or optimal control problems are omnipresent in the (bio)chemical industry. In addition, these problems often involve multiple and conflicting objectives, leading to a so-called set of Pareto optimal solutions, instead of one single optimum. Alternatively, robustness of the obtained solutions with respect to model uncertainties, e.g., guaranteeing that critical constraints are not violated, is of the highest importance in process industry. Moreover, robustness can also be interpreted as an additional and conflicting objective, since more robust solutions typically induce a performance decrease. The current manuscript exploits advanced deterministic techniques to efficiently and accurately generate Pareto sets in the presence of model uncertainty. The developed procedures allow the presentation of robust Pareto sets, i.e., Pareto sets in which robustness is an additional objective. Based on these Pareto sets, all trade-offs can clearly be assessed by the decision maker. Two illustrative case studies are presented for the optimal design and operation of a jacketed tubular reactor with conflicting conversion and energy objectives and a fed-batch bioreactor with conflicting productivity and yield aims. © 2011 Elsevier Ltd.status: publishe