First-order rule partitions-based decomposition technique of type-1 and interval type-2 rule-based fuzzy systems for computational and memory efficiency

Rule-based fuzzy systems are widely adopted in various fields for their effectiveness in reducing uncertainties. However, large rule bases present significant computational and memory challenges, especially for real-time applications. To overcome this limitation, this paper introduces a novel first-order rule partitions-based decomposition technique (FORPs-DT) for type-1 (T1) and interval type-2 (IT2) fuzzy logic systems (FLSs). This method involves decomposing the universe of discourse, membership functions (MFs), and fuzzy sets (FSs) by defining conditions for nonzero firing levels (intervals). Significantly, this approach allows the construction of subfuzzy systems with separate knowledge bases while maintaining the input-output relationship. The paper provides detailed explanations and visual representations of FORPs-DT and proposes a case of identical FORPs to illustrate its simplicity and resource optimization benefits. Extensive experiments demonstrate the superiority of this technique, showcasing not only reduced execution time and memory usage but also enhanced performance through identical partitions, which allows an increase in the number of FSs without additional computational or memory overhead.