| Semiconductors represent a large and growing high tech industry and are incorporated in a growing array of products. Long and complex semiconductor processes make planning for semiconductor manufacturing challenging; embedding semiconductors in products involves similar challenges for those supply chains. Models for yields and cycle times of semiconductors and supply chains with semiconductors were used to provide an approach for decisions about quantities and timings for starting material as well as co-optimization of on-time delivery, quoted lead time and acceptable inventory levels.; The probability of meeting on-time delivery expectations for semiconductors was modeled as the product of the probability that sufficient product will be available for delivery, allowing for variability in yield, and the probability that the product will be shipped in the committed timeframe, allowing for cycle time variance. The on time delivery model for semiconductors can forecast the on time delivery probability and likely inventory tradeoffs given inputs on the orders for semiconductors, delivery date or lead time expectations, existing inventories, lot size decisions in wafer fabrication, expected yields and expected cycle times. Semi-empirical modeling of yield distributions for semiconductors used a defect density model to obtain relationships between the mean and the standard deviation of probe yield distributions. Semi-empirical modeling of cycle time distributions allowed distributions of cycle times to be estimated from expected cycle times. The semiconductor on time delivery model was applied in setting recommendations for inventory levels for families of integrated circuits, setting acceptance limits for probe yields and cycle time expectations for planners at a semiconductor company.; The statistical model for supply chains with semiconductors was based on a network of parallel semiconductor components that converge at module assembly; component involved a series of stages representing semiconductor manufacturing. The model forecasts the lead time and inventory cost for the network, provides a mean for evaluating the lead time, on time delivery, inventory level tradeoffs and provides recommendations on inventory levels for the supply chain. |
