Supply Chain Management

Control-oriented Approaches to Supply Chain Management and Enterprise Systems:

Enterprise systems management in general, and supply chain integration in particular, has become a strategic imperative for most manufacturers. A "supply chain" (also known as a demand network) consists of interconnected components required to transform ideas into delivered products and services. Supply chain integration applies a total systems approach to managing the entire flow of information and materials, from the raw materials suppliers through the factory and distribution centers to the customer. Traditionally, supply chain strategies have focused on integrating functions within a single company to provide a context for more effective decisions. Today, industry leaders recognize that the success of their business strategies also depends on closer integration with their core suppliers, distributors, and other business partners.

Our research in this area (funded by an NSF grant CMMI-0432439 and the Intel Research Council) has been to develop control-oriented approaches to managing supply chains using decision policies based on Model Predictive Control (MPC).  The work has explored both the development of MPC policies contextualized for high-volume discrete-parts manufacturing and the associated problem of generating control-relevant demand forecast signals tailored for these decision policies.  The collaboration has been highly interdisciplinary in nature, involving faculty from the Department of Mathematics (Armbruster, Kawski, Mittelmann and Ringhofer), Computer Science (Sarjoughian) with research scientists from Intel Corporation on formulations meaningful to tactical inventory planning in semiconductor manufacturing. Our solutions are evaluated on a series of benchmark problems developed by Intel Corporation which involve multiple wafer fabs, assembly and test facilities, distribution warehouses, and customer locations.