Abstract
Forecasting and production control systems typically rely on operational rules that have been accumulated and refined from enterprise experts. Designing a rule-based system is a challenging task. In this article, a new rule-based system design methodology for forecasting and production control is proposed. The methodology first represents the rule-based system as a finite state automaton (a Moore machine) and then formulates an optimal control problem in a feedback control architecture. The solution to the optimal control problem provides action rules for forecasting and production that minimize cost over a given time horizon. The proposed methodology provides a systematic tool for rule-based system design that gives robust and realistic solutions.
Acknowledgement
This research is sponsored in part by the National Science Foundation through Grant CMMI-0908317.