Analytical Model for Production Planning in Integrated Circuit Supply Chains

摘要

Current production planning models do not account for binning in integrated circuit (IC) manufacturing. Binning is the process by which ICs are sorted into categories based on their performance and using the different bins of ICs for different products. In production planning, the goal is to determine the quantity of a product to produce, but this is difficult in IC supply chains due to the inherently high variability in both yield and binning rates. Typically, supply chains must account for variation in demand, but in IC supply chains must also account for variation in supply. The most common production planning method, material requirements planning (MRP), can't account for this variability, resulting in either frequent stockouts or excess inventory. In this paper, we develop two models that account for this variability. We use discrete event simulation based on the Dirichlet distribution, and an analytical model based on the beta-binomial distribution. These two models accurately model the uncertainty in supply due to yield and binning variability, but discrete event simulation has the flexibility to model a wider variety of supply chains. Further research is necessary to validate these models in a real-world supply chain.