In this study, a dispatching rule based genetic algorithm with fuzzy fitness value is proposed to provide acceptable solutions with respect to multiple performance measures. The objectives considered are makespan, average flow time, number of tardy jobs, and total tardiness. In the proposed genetic algorithm, dispatching rules for each machine in different time intervals are encoded in the chromosome and then a simulation model is constructed to determine the values of these performance measures. A two-level fuzzy structure is used to calculate the fitness value. In the lower level, satisfaction levels for individual performance measures are determined by using fuzzy membership functions. In the higher level, an overall satisfaction level is obtained by considering all performance measures and this value is used as the fitness value of the chromosome. Later, fuzzy membership shapes and fuzzy operators are investigated, and the results are compared with non-dominated genetic algorithm. The results show that the proposed approach can quickly capture better schedules that highly satisfy decision makers.
展开▼
