Revamping America’s Freight System: A Smart Approach to Chaos
America’s freight system stands as a modern marvel of logistics, efficiently moving goods across vast distances. Yet it remains vulnerable to disruptions—be it a fierce snowstorm or a simple truck delay outside a major hub. When these incidents occur, they expose the fragility of a system built on the assumption of predictability, revealing how freight operates amidst chaos.
Enter Dr. Lacy Greening, an assistant professor in industrial engineering at Arizona State University. Recently named one of 15 semifinalists in the U.S. Department of Transportation’s ARPA-I Innovation Challenge, Greening is poised to tackle this complex issue head-on.
Understanding the Middle Mile
Greening’s innovative concept is rooted in a straightforward question: How can the U.S. freight system become smarter and more resilient through adaptive, real-time networks?
Currently, the freight logistics landscape is hindered by fragmented planning systems. When one component malfunctions, people scramble to repair the entire operation manually.
“The existing planning is often sequential,” Greening explains. “We conduct route planning, dock scheduling, and sortation independently. A delay in one area requires manual intervention to rectify all downstream effects.”
Her proposed solution involves “agentic AI,” which harnesses multiple smaller AI agents that coordinate locally while simultaneously communicating with the broader system. This structure allows for real-time adjustments across the freight network, eliminating the need for human intervention in many instances.
“Solving the entire problem at once is impractical. Our aspiration is to have dedicated models that communicate seamlessly,” Greening notes.
Greening’s work is complemented by Reem Khir from Purdue University, focusing specifically on the freight pipeline’s most intricate and often costly segment: the middle mile. This stage encompasses all transfers between warehouses, fulfillment centers, and regional hubs, bridging the initial delivery from factories and ports to the final delivery to consumer doorsteps.
“The middle mile tends to be the most complex portion of the system,” Greening adds. “It involves substantial consolidation and incurs significant costs.”
Current tools are insufficient for managing middle-mile operations, which typically respond reactively to disruptions. When issues arise, it can take from 30 minutes to an hour for manual responses to implement, by which point delays may have already spread.
Reducing Disruption Costs
The ARPA-I Innovation Challenge is designed to challenge existing norms and promote transformative technologies aimed at enhancing the safety, reliability, and cost efficiency of America’s infrastructure.
Greening’s proposal outlines a three-tiered agentic AI framework. Her approach begins with AI agents collecting and analyzing real-time data, including traffic updates, weather forecasts, equipment status, and workforce availability. In the next layer, planning agents utilize optimization and machine learning techniques to adjust truck routes, reschedule dock allocations, and rebalance resources dynamically. The top layer comprises human oversight, ensuring decision validation and intervention when necessary.
This system could function like a freight network with inherent reflexes, capable of preemptively addressing issues before they escalate.
“For instance, if a massive snowstorm is imminent, the system could preemptively reroute trucks,” Greening illustrates. “Our goal is to prevent disruptions instead of merely reacting to them.”
By focusing on prevention, not only does this initiative promise speed and responsiveness, but it also aims to cut costs associated with delays—costs meaning companies could avoid resorting to pricier alternatives like air freight, team drivers, or overtime labor.
“Delays create problems, and those problems inflate costs,” Greening emphasizes.
Having progressed to her next phase as an ARPA-I semifinalist, Greening recently participated in a U.S. Department of Transportation Innovation Workshop in Washington, D.C., presenting her concept and garnering insights from experts across government and industry sectors.
A Promising Career Trajectory
Greening’s recognition comes early in her academic journey but is rooted in extensive industry experience. She earned her Ph.D. from the Georgia Institute of Technology, where she also served as a Dwight D. Eisenhower Transportation Fellow, collaborated with The Home Depot, and interned with Amazon on middle-mile planning and optimization.
“Industrial engineers have the opportunity to resolve complex issues across various industries,” she reflects. “The reality of supply chains particularly resonates since the complications are tangible and carry real-world implications.”
Ross Maciejewski, director of the School of Computing and Augmented Intelligence, commends Greening’s work for underscoring the need for enhanced systems-level thinking in transportation, particularly with the rising prominence of AI in decision-making.
“Lacy’s research embodies the innovative spirit that ARPA-I aims to cultivate,” Maciejewski states. “She isn’t just applying AI to transportation; she’s reimagining how complex infrastructure systems can detect disruptions, adapt in real-time, and ultimately enhance efficiency at scale.”
As supply chains remain under pressure and interruptions become increasingly common, Greening’s inquiry poses an essential question: What if freight systems could learn from chaos rather than merely withstand it?
The answer could play a crucial role in shaping the resilience of America’s infrastructure.
This story originally appeared on ASU’s School of Computing and Augmented Intelligence website. Explore more stories about the program here.
