Run-to-run Optimization via Control of Generalized Constraints

Run-to-run optimization methodologies exploit the repetitive nature of batch processes to determine the optimal operating policy in the presence of uncertainty. In this paper, a parsimonious parameterization of the inputs is used and the decision variables of the parameterization are updated on a run-to-run basis using a feedback control scheme which tracks signals that are invariant under uncertainty. In this run-to-run framework, terminal constraints of the optimization problem and cost sensitivities constitute the invariant signals. The methodology is adapted to improve the cost function from batch to batch without constraint violation.