Evolutionary algorithm based fuzzy modeling using conciseness measure

In this paper a fuzzy modeling method using a new conciseness measure is presented. Conciseness of fuzzy models is defined by the shape and allocation of membership functions and the conciseness is quantified by introducing fuzzy entropy. This paper proposes a new measure which evaluates the deviation of a membership function from symmetry. The measure has a different aspect from De Luca and Termini´s (1972) fuzzy entropy measure, which could only evaluate the shape of a membership function. By combining these two measures, a, new measure is derived for evaluation of the shape and allocation of membership functions of a fuzzy model. Numerical results show that the new conciseness measure is effective for fuzzy modeling formulated as a multi-optimization problem