Demand Management in Semiconductor Manufacturing: A Dynamic Pricing Approach Based on Fast Model Predictive Control

Abstract

Demand Management (DM) is a method which can help a firm to sell the right inventory unit to the right type of customer, at the right time, and for the right price. DM guides the decision of how to allocate undifferentiated units of capacity to available demand in such a way to maximize revenue. The problem then becomes one of determining how much to sell at what price and to which market segment. The objective of this research project is to present control system approaches to DM with Dynamic Pricing using Fast Model Predictive Control in the Semiconductor Manufacturing arena. Manufacturers of integrated circuits invest billions of dollars in process equipment, and they are interested in obtaining as rapid a return on their investment as possible. Rapid yield learning is thus becoming an increasingly important source of competitive advantage. The sooner a potentially lucrative circuit yields, the sooner the manufacturer can generate a revenue stream.