FUZZY SET SHAPE OPTIMIZATION

This paper deals with design and implementation of fuzzy set shape optimization utilizing general optimization framework based on evolution algorithms for improving accuracy of fuzzy approximation employing Mamdami and Assilian method. Fuzzy sets are represented by trapezoids entered by the user and should approximate given real function. Approximated function is given in the form of discrete points, density of which can vary. The optimization framework that is exploited is designed in layers to truly depicts both the model and application domains. The framework itself can be used for broad collection of evolutionary algorithms, the solutions of which are based on populations of possible solutions. The paper presents design of the application layer classes, their relations thus specifying concrete usage of the framework, achieved results and future research.