Handling input dynamics in integrated scheduling and control

Coordinating production scheduling decisions with the process control system requires considering the evolution of the process over multiple time scales and at multiple levels of detail. From a mathematical perspective, this requires dealing with process models that are large-scale, ill-conditioned and involve both continuous and discrete variables (the former related to physical states, while the latter reflect production management decisions). In this paper, we introduce a novel methodology for time scale-bridging between production scheduling and process control. We use process operating data to obtain low-order models of the closed-loop behavior of the process, which are then incorporated in the production scheduling framework. The theoretical developments are accompanied by an illustrative case study on a methyl methacrylate process, showing excellent economic results and significantly improved computational performance.