Abstract
The thin-film transistor-liquid crystal display (TFT-LCD) array production is usually the longest process of TFT-LCD production, which is critical to maximize the throughput and satisfy the demand. Among the array process, the complex constraints should be considered in different workstations, such as the photo mask availability in photolithography station, and the setting of setup time in deposition stations. To increase the throughput as well as decrease the cycle time, tardy jobs and total tardiness, this study developed a simulation-based scheduling and dispatching system (SSDS) for array manufacturing process with the considering of the complex constraints in different stations. A hybrid genetic algorithm (HGA) is used to maximize the utilization of the bottleneck station, while selected dispatching rules were employed for non-bottleneck stations to cooperate with the scheduling of bottleneck station. To validate the proposed approach, the experiments based on empirical data were tested and indicate the proposed HGA outperforms the conventional approaches for photo scheduling problem. The implementation of selected dispatching rules based on the performance evaluation are also able to increase throughput and decrease cycle time, tardy jobs and total tardiness.