Correlated multi-mode disruptions raise freight costs over 25% beyond historical estimates.

Freight networks move goods using multiple transport modes, like trains and barges. When bad weather hits, several routes can fail together. Traditional safety checks usually look at past events or assume problems happen independently, which underestimates real dangers. To fix this, researchers created a new stress-testing framework. They used generative adversarial networks, a type of artificial intelligence, to create realistic disruption scenarios. Then, an optimization model rerouted shipments to calculate system costs across the Tennessee Valley corridor. They found that historical disruptions increased total cost by about 3%. In contrast, the AI-generated scenarios increased costs by over 25%, reaching an expected annual cost of $5.11 million. Risk heavily concentrates in correlated multi-node failures and key hubs like the Port of Knoxville.

Why it matters

Freight operators and regional planners currently base resilience plans on past disruptions, which underestimate the cost of linked failures across trucks, trains, and barges. This framework shows that such compound events can raise costs by over 25%, meaning millions of dollars in unanticipated losses. It pinpoints the specific nodes and failure combinations that drive those costs, letting agencies invest in protection where it matters most.

Questions answered