Uncertainty versus choice in rule-based fuzzy logic systems

Abstract

We demonstrate that some of the so-called uncertainties that may be present in a rule-based fuzzy logic system (FLS) are not uncertainties, but are instead choices that must be made as a result of the rich variety of mathematical models associated with the various elements of a FLS. We have established a hierarchy for the uncertainties and the choices, one that will hopefully guide us in developing FLS models that can better account for all sources of uncertainties than do present models. It seems that, at the very least, we must account for the uncertainties present in all rule-words, including connector words. This can be accomplished by using type-2 fuzzy sets for antecedent and consequent words and parametric operators for connector words.