Process Optimization and Control under Chance Constraints

We propose to use chance constrained programming for process optimization and control under uncertainty. The stochastic property of the uncertainties is included in the problem formulation. The output constraints are to be ensured with a predefined confidence level. The problem is then transformed to an equivalent deterministic NLP problem. The solution of the problem has the feature of prediction, robustness and being closed-loop. In this paper, the basic concepts and solution strategies are discussed to illustrate the potential for optimization and control under uncertainty.