An inventory planning methodology for a semiconductor manufacturer with significant sources of variability

Abstract

The research and thesis focus on the development and implementation of a data-driven inventory and production planning methodology for the Eastman Kodak Company's Microelectronics Technology Division (MID). The thesis gives an introduction to MID's products, processes, and customers, as well as the current business conditions, recent improvements, and challenges present at MID at the time of the research. The thesis examines the major causes of the need to hold inventory: to buffer against low yield, yield variability, long cycle times, cycle time variability, and uncertain demand. The thesis builds upon extensions of the Base Stock model to recommend target finished goods inventory (FGI) and work-in-process (WIP) levels, as well as material release decisions. The model also allows MID to predict expected customer service levels under their actual inventory situation. While the methodology is demonstrated for one important MTD product, the methodology can potentially be used for other MTD products with similar regular demand patterns. While holding inventory provides short-term solutions to problems such as cycle time variability, the framework provides a means for quantifying the impact of production system improvements on customer service and inventory costs. The model allows MTD to estimate the inventory and cost savings that would result from longer-term improvements such as cycle time reduction, yield improvement, or variability reduction. The thesis also outlines the processes for using this framework on an ongoing basis, periodically adjusting for changes to production system performance measures or demand.