A soft computing approach to mastering paper machines

Paper machines are extremely complex systems which require a deep knowledge of the relations between levels of usage of factors and characteristics of the process/final product. In this paper we propose to capture the tacit knowledge of experts in the form of a fuzzy ontology. Based on the fuzzy ontology, we introduce a knowledgebased system which can ultimately recommend what factors should be increased, or decreased, in order to obtain a different and better output. The system is based on an integer linear goal-programming optimization problem whose parameters come from the fuzzy ontology. We also propose extensions of our model to account for additional constraints and knowledge expressed in the form of fuzzy numbers. © 2012 IEEE.