Rule-driven manufacturing control based on ontologies

Manufacturing control is an increasingly challenging task in many domains due to the complexity of production processes. In many Manufacturing Execution Systems, rule-based components therefore steadily become harder to analyze and to maintain as the number of rules usually grows with the age of the factory. As a result, given large rule sets accumulating over the years of operation, the behavior of such systems is hard to predict and changes to rule sets or referenced master data may lead to undesired and often very costly behavior. This paper describes an approach to rule-driven decisions based on knowledge about the factory stored in ontologies. Expressing rules and master data in OWL and reasoning on these knowledge bases allows better understanding of the control systems and more intuitive rules. Most importantly, several features of the rule sets and the corresponding master data can be verified whenever changes are necessary thus leading to a more predictable control system even with large rule sets.