Recent Research Projects


Network Optimization for a Confectionary Supply Chain (with Just Born)

Image of a Yellow Marshmellow PeepJust Born is the manufacturer of Peeps, Mike and Ike, Hot Tamales, and several other well known candy brands. We constructed a model to determine (1) which pool points (distribution centers) Just Born should use nationwide, (2) which customers should be served by pool points and which should be shipped to directly, and (3) how trucks should be routed through the network. We developed an integer programming (IP) model that is also general enough to apply to supply chains for other companies and industries. We embedded this model into a user-friendly graphical user interface (GUI) that runs on-site. In addition to generating substantial freight savings, our model also helped Just Born develop a multi-company freight-consolidation effort known as the "Confection Connection."

The project was described in several media reports, including:

This project was financed (in part) by a grant from the Commonwealth of Pennsylvania, Department of Community and Economic Development, through the Pennsylvania Infrastructure Technology Alliance (PITA).



Capacity Planning for a Gases Supply Chain with Power Disruptions (with Air Products)

Picture of AIR Products tanker at plantThis project developed a mathematical optimization model for the design and management of a supply chain network that uses inventory as a tool for buffering against power disruptions. The model is applicable for firms that operate under Interruptible Load Contracts (ILC), in which the electric utility may shut off power to factories during periods of peak demand in exchange for lower electricity rates. Our model optimizes production and inventory levels while ensuring adequate service to all customers during disruptions, even with multiple unpredicted interruptions, using robust optimization methodology. It can be used as an operational tool to set production and inventory levels, and as a strategic tool to help managers evaluate the financial impact of agreeing to an ILC. Although the model was designed with the liquefied gases industry in mind, was general enough to apply to similar infrastructures in other industries.

This project was financed (in part) by a grant from the Commonwealth of Pennsylvania, Department of Community and Economic Development, through the Pennsylvania Infrastructure Technology Alliance (PITA).



Robust Network Design for Distribution under the Threat of Disruptions (with Air Products)

Picture of AIR Products plant at night.We developed a strategic model for supply chain network design under the threat of disruptions. Given the hazardous, volatile nature of some chemical products, infrastructure reliability is critical for chemical companies, as well as for the states and countries in which the infrastructure operates. When disruptions occur, material must be re-routed through the network. Our integer programming (IP) model chooses which nodes and arcs to open to minimize the nominal cost (the cost if no disruptions occur) while also ensuring that the cost in any disruption scenario is bounded by a specified level. The model can be used to evaluate the tradeoff between cost and robustness. Our particular approach for modeling robustness (called “p-robustness”) accounts for decision-makers' risk aversion, a particularly important consideration for disruption models since disruptions are low-probability, high-impact events that may contribute little to the nominal cost but may be disastrous if they occur.

This project was financed (in part) by a grant from the Commonwealth of Pennsylvania, Department of Community and Economic Development, through the Pennsylvania Infrastructure Technology Alliance (PITA).



Managing Capacity and Inventory in the High-Tech Manufacturing Industry (with LSI)

Picture of computer componentWe developed a quantitative model for managing capacity and inventory in the high-tech manufacturing industry. Our model considers a semiconductor supply chain with four echelons (stages) and multiple products. It develops answers to three main questions: (1) Whether a given product should be produced based on target inventory levels or target production rates; (2) How much capacity (internal and external) to have over time; (3) Where in the four-stage process to have inventory buffers. The model uses linear programming (LP) to answer these questions, with the objective of maximizing gross margin. It also provides a detailed plan for production and inventory levels to serve as a starting point for the tactical planning process. The model was developed and implemented in part by Lehigh Ph.D. students as part of their dissertation research.

This project was financed (in part) by a grant from the Commonwealth of Pennsylvania, Department of Community and Economic Development, through the Pennsylvania Infrastructure Technology Alliance (PITA).



Strategies for Coping with Supply Chain Disruptions (with IBM Integrated Supply Chain)

Picture of IMB Data Center.In this project, we formulated analytical models that examine strategies for coping with supply chain disruptions when the firm has advanced warning of a disruption, for example, an approaching hurricane or heightened terror alert status. The firm updates its “threat level” and reacts by increasing or repositioning its inventory. We formulated an inventory model and a novel heuristic technique for finding near-optimal inventory levels in a multi-echelon supply chain as the threat level changes.

(Photo credit: Courtesy of International Business Machines Corporation. Unauthorized use not permitted.)