The three-year project “Enhancing Stochastic Supply Chains via Cascading Simulation and Adjustable Optimization” is supported through NSF’s Division of Civil, Mechanical, and Manufacturing Innovation.
Ningji Wei, an assistant professor in the Department of Industrial, Manufacturing & Systems Engineering (IMSE) at the Edward E. Whitacre Jr. College of Engineering (WCOE) received a $315,912 grant to evaluate variables in supply chains that can create
vulnerabilities and to develop strategies that minimize negative impacts.
Supply chains are complex systems of many organizations that are connected in a network
of multi-level interactions and exchange of goods and services. Any organization in
the network may know very little about other organizations in the chain if they are
not directly interacting upstream or downstream.
These systems face uncertainty from a variety of internal and external factors that can include suppliers not able to produce the quantity of goods needed, changes in demand, transportation issues, or even impacts from natural disasters or other situations like COVID, which can affect multiple components of the chain in different ways at the same time.
The purpose of Wei’s project is to develop a process that considers all of these elements together as a system and will provide stakeholders with a data-driven, timely way to respond to these uncertainties so that negative impacts are minimized.
“This NSF award is a tremendous achievement for Dr. Wei and a strong reflection of the caliber of research being conducted in our department. His work on resilient, data-driven supply chains addresses challenges with real consequences for industry and society, and we are excited to see the impact it will have,” said Burak Eksioglu, department chair of IMSE.
The project will roll out in three stages.
The research will draw on Wei’s expertise in decision-making under uncertainty and his co-investigator Peter Zhang, an assistant professor of operational research in Heinz College at Carnegie Mellon University, whose focus is supply chain management. In the first stage, Wei and Zhang will map the system using existing information.
“It’s important to understand, though, that for any supply chain, a single organization is unlikely to have complete information. That is very common in supply chain management. This problem of opacity is one of the primary challenges that creates disruptions in supply. It’s hard to know what to fix if you can’t identify where the problem may be,” Wei explained.
Because it is not possible to know every detail of every organization (called a node) and every upstream or downstream interaction (called an edge), Wei and Zhang will apply graph theory and network optimization techniques to infer the missing information and create a mathematical model of the system. This model will provide a way to evaluate the probability of system interactions, giving stakeholders a measure of confidence in decision-making, even in circumstances where there may be gaps in information.
During the second stage of the project, Wei will utilize simulation and stochastic optimization techniques to study a variety of uncertainties (hurricane, transportation issue, etc.) to determine the impact to system performance. The simulations will look at how specific adjustments at a node impact the entire system’s performance to determine which adjustments maintain or improve supply during that uncertainty event.
“This stage is about identifying potential vulnerabilities in the system related to each uncertainty event so that when that event occurs, decision makers are able to more easily implement the correct adjustment that will stabilize the system’s performance,” Wei added.
“On something as complex as a supply chain, knowing how to react efficiently in response to new information and apply adjustments at the correct nodes, or the coordination between several nodes to minimize disruptions, is crucial.”
In the final stage, they expect to develop a mechanism so that supply chain management professionals are not limited to just looking at one decision ahead of them, they have a tool that enables them to launch adaptive solutions in real time as new information is added to the model.
The project funds research support for two doctoral students. Wei will incorporate component elements of the project undergraduate coursework and high school-level activities so that students gain experience in mathematical modeling and simulation and learn how to apply those approaches to identifying and solving real-world problems.