Abstract
This paper proposes a complexity management methodology for fuzzy systems with feedforward rule bases. The methodology is based on formal methods for presentation, manipulation and transformation of fuzzy rule bases. First, Boolean matrices are used for formal presentation of rule bases. Then, binary merging operations are used for formal manipulation of rule bases. Finally, repetitive merging operations are used for formal transformation of rule bases. The formal methods facilitate the understanding and modelling of fuzzy systems in terms of interacting subsystems. In particular, the methods reduce the qualitative complexity in fuzzy systems by improving the transparency of the rule bases.