Solving Fuzzy Job-Shop Scheduling Problems with a Multiobjective Optimizer

In real-world manufacturing environments, it is common to face a job-shop scheduling problem (JSP) with uncertainty. Among different sources of uncertainty, processing times uncertainty is the most common. In this paper, we investigate the use of a multiobjective genetic algorithm to address JSPs with uncertain durations. Uncertain durations in a JSP are expressed by means of triangular fuzzy numbers (TFNs). Instead of using expected values as in other work, we consider all vertices of the TFN representing the overall completion time. As a consequence, the proposed approach tries to obtain a schedule that optimizes the three component scheduling problems (corresponding to the lowest, most probable, and largest durations) all at the same time. In order to verify the quality of solutions found by the proposed approach, an experimental study was carried out across different benchmark instances. In all experiments, comparisons with previous approaches that are based on a single-objective genetic algorithm were also performed.