The impact of lot splitting on due date performance

The primary purpose of the research project was to determine if the flow time reductions that occur from splitting jobs into smaller transfer batches result in improved due data performance. The research was performed using a computer simulation of a five machine open job shop. In the open shop environment, each job is unique, and the flow time reductions are solely a result of overlapping operations rather than saved setups from processing similar jobs in sequence. To determine the impact of lot splitting on due date performance, two questions had to be answered. First, how does the impact of lot splitting vary with the magnitude of setup times? Second, how does the size of the transfer batch affect lot splitting performance? This study shows the benefits from lot splitting improve monotonically with decreasing transfer batch size and that lot splitting provides improvements that are essentially independent of the setup time as long as the total utilization (ratio of setup and processing time to available time) is constant. As for the primary research question, lot splitting can provide substantial reductions in mean tardiness (up to 39%) and number of jobs tardy in the open shop environment. The repetitive lots logic developed by Jacobs and Bragg (1988) is reformulated to clearly show that it can be used in combination with any dispatching rule. This rule demonstrates that the repetitive lots logic makes surprisingly good scheduling decisions in spite of its simplicity. Lot splitting is shown to be a simple and effective means of improving the performance of an open job shop. Based on its robust performance in this study, the benefits should be able to be realized in combination with other shop floor control techniques, such as order review release and order expediting.