Managing variability in the semiconductor supply chain

Abstract

Within the semiconductor industry, the variability in both supply and demand is quite high; this uncertainty makes supply chain planning very difficult. We analyze the current tools and processes at a large semiconductor manufacturing company and then propose a framework for improvement based on hierarchical production planning. We present an appropriate decomposition for this specific planning problem and illustrate some limitations of traditional inventory models. New safety stock equations are developed for this planning problem based on a simple analysis using the basic ideas from probability theory. We also devise a new method to determine lead times that more accurately captures the actual lead time seen in the supply chain. Finally, an algorithm is developed to determine appropriate inventory levels and production allocation. These ideas, when used together, provide a powerful framework to properly manage supply chains in highly stochastic environments.