A methodology for improving on-time delivery and load leveling starts

Abstract

We will describe the procedure and implementation of an iterative simulation-based production scheduling and planning system. Beginning with the total demand and initial work-in-process (WIP) status, the system can efficiently generate a capacity-feasible, load-leveled start schedule that results in high resource utilization, a minimum number of late orders and reduced labor variability. There are two phases in the system. In phase one, the system utilizes a demand pre-processor to aggregate backlog, safety stock and forecasted demands by priority and by period. These demands, including those that are defined against product families and groups, are then exploded down to the part number level. Utilizing heuristic algorithms, the system next determines start dates relative to capacity constraints, order due dates and order priorities. In phase two, the system again adjusts the start time of each lot so that resources are exploited at higher utilization and the production load is leveled regardless of variability in the demands. We will present simulation-based algorithms specifically designed for each phase of the system. The algorithm for phase one is executed to achieve the best on-time delivery performance by using the estimated cycle times from the previous iteration sequences. We also discuss an alternative for terminating the successive iterations based on the rate of improvement in overall lateness. Phase two takes the lot start times from the first phase and adjusts them in order to compensate for capacity constraints as demand for future periods is added.