Harmonizing the use of optimization and feedback in process operations and control.

This paper addresses how to combine the two main engineering design tools, optimization and feedback, to obtain high performance process control or process operation systems. We discuss the historical development of process control, where feedback dominated the early practical designs, and optimization was then incorporated much later through technologies like model predictive control. In chemical production scheduling, on the other hand, optimization had a strong early influence on the problem formulation, and feedback has only recently made an appearance. The recent developments in both process control and scheduling are illustrated with specific examples emerging from this series of FOCAPO/CPC meetings. The paper next presents recent theoretical developments in nominal and stochastic model predictive control. The closed-loop properties that arise from these different open-loop optimal control problems are then compared. The paper closes with some discussion of when the improvements of the closed-loop properties are worth the added complexity of the stochastic optimal control problem. • Paper addresses how to combine optimization and feedback to obtain high performance process control systems. • Illustrates recent developments with examples from this series of FOCAPO/CPC meetings. • Presents recent theoretical developments in nominal and stochastic model predictive control. • Discusses when improvements in closed-loop properties are worth the complexity of a stochastic model. [ABSTRACT FROM AUTHOR]