Abstract
Extrusion of rubber profiles is a time- and resource-intensive process. The experience of individuals supervising the production is difficult to gather and to convey to others. The aim of the MErKoFer project is to gather accumulated knowledge on the production process and established procedures, which are employed in handling occurring malfunctions and ensuring specified quality standards, and to give advisory feedback based on this knowledge. This approach is based on case-based reasoning, an ontological representation of domain-specific information and dependencies, associations and rules identified by data mining techniques, and neural networks. Solutions of similar situations are provided to the machine operators as hints from the knowledge base to support their decisions and improve their autonomy and the quality of their actions.