2nd International Conference on Modelling, Computation and Optimization in Information Systems and Management Sciences, Metz, Fransa, 8 - 10 Eylül 2008, cilt.14, ss.32-42
Primary objective of this study is to show fuzzy optimization models can be solved directly by employing metaheuristics and ranking methods without requirng a transformation into a crisp model. In this study, a fuzzy multi-item Economic Order Quantity (ECQ) model with two constraints is sovled directly (without any transformation process) by employing three different fuzzy ranking functions and the Particle Swarm Optimization (PSO) metaheuristic. The parameters of the problem are defined as symmetric triangular fuzzy numbers. Having fuzzy paramers, the objective function values of the generated soution vectors also will be fuzzy numbers. Therefore, in the selection of the best solution vector, ranking of fuzzy numbers is used. Similarly, the feasibility of the constraints for the generated solution vectors will be determined via ranking of two fuzzy numbers. By this approach other fuzzy optimization problems can be sovled without any transformation process.