This paper presents a genetic algorithm (GA) approach to solving fuzzy constraints linear programming (FCLP) problem. We use GAs to solve this fuzzy problem without defining the membership function for fuzzy numbers or using a penalty method for constraint violations. The proposed approach simulates every fuzzy number by distributing it into certain partition points. The final values obtained after the evolutionary process represent the membership grade of that fuzzy number. The computation of fuzzy equations by GAs does not require the conventional extension principle or interval arithmetic and alpha -cuts for solving FCLP. Instead, GAs use the usual evolutionary process. The empirical results show that the proposed approach obtains very good solutions within the given bounds of each fuzzy coefficient compared with other fuzzy methods. The fuzzy concept of the GA approach is different but gives better results than other traditional fuzzy methods.
關聯:
INTERNATIONAL JOURNAL OF INNOVATIVE COMPUTING INFORMATION AND CONTROL
