As shown in Fig. 1, a FLS contains four main modules. First, fuzzi-fication module that transforms the crisp numbers inputs into fuzzy sets. This is by using membership functions. Second, knowledge base stores the IF-THEN rules. Third, inference engine is used to establish fuzzy conclusions. Finally, DEFuzzification module transforms the obtained fuzzy conclusion into a crisp value.
Mamdani and Takagi Sugeno (TSK) are the most utilized FLS models in the literature. They mainly differ in the output structure. In Mamdani systems, both inputs and outputs are fuzzy propositions using linguistic variables. In TSK models, system out-puts are numerical values rather than linguistic variables. The outputs can be constants, polynomials, functions or differential equations.