ON A HIERARCHICAL STRUCTURE FOR FUZZY MODELING AND CONTROL

The fuzzy logic controller is examined and the basic assumptions inherent in the Mamdani model is described. Distinctions are made between rule firing based upon possibility and certainty qualification. Rules am looked at as a partitioning of the input space. We discuss the use of certainity qualification in determining the firing level of a rule. Different representations of the rule consequent are discussed. A new structure for the representation of rules in fuzzy systems is introduced that is called the hierarchical prioritized structure (HPS). This new HPS in addition to providing a useful structure for representing knowledge allows for a natural framework for learning rules.