Multi-objective evolutionary design of fuzzy rule-based systems

This paper clearly demonstrates advantages of our evolutionary multiobjective optimization approach to the design of fuzzy rule-based classification systems over single-objective methods. The main advantage of our approach is that a large number of tradeoff (i.e., nondominated) fuzzy rule-based systems can be obtained by its single run with respect to conflicting objectives: accuracy maximization and complexity minimization. By analyzing the obtained fuzzy rule-based systems, the decision maker can understand the tradeoff between these two objectives. Such understanding is of great help when the decision maker chooses a final compromise fuzzy rule-based system. In the case of single-objective methods, only a single fuzzy rule-based system is obtained based on the pre-specified preference of the decision maker. We compare four formulations of genetic algorithm-based rule selection through computational experiments on well-known benchmark data sets. The four formulations have two objectives, their weighted sum, three objectives, and their weighted sum, respectively.