Abstract
Hurricane evacuations face a critical paradox: the very vehicles intended to save lives often create paralyzing gridlock, leaving the most vulnerable populations stranded. Traditional evacuation plans are primarily based on private automobiles, overlooking the strategic potential of public transit and the dynamic nature of traffic congestion generated during mass departures. This paper introduces a comprehensive optimization framework that elevates public transit from a peripheral support role to a central tool for managing congestion and promoting equitable access. The proposed framework integrates a rolling-horizon column generation model that accommodates multi-trip bus operations over a 72 h evacuation window, a dynamic traffic equilibrium component that updates travel times based on cumulative network congestion, and equity constraints that ensure fair regional coverage. Applied to Charleston County, South Carolina, U.S., the model exposes the risks of static planning: conventional approaches, which ignore real-time congestion feedback, produce infeasible plans that would lead to over 16,000 individuals being unsafely delayed, while our dynamic model successfully evacuates 99.6% of transit-dependent residents. Equity analysis quantifies a “price of fairness,” revealing trade-offs between efficiency and regional balance, and a sensitivity experiment with explicit shelter capacity constraints illustrates how inadequate shelter space can become the binding bottleneck even when road capacity is sufficient. Additionally, this paper identifies a critical system tipping point at 15%–20% public transit dependency, beyond which network performance declines sharply. The computational efficiency of the framework enables real-time implementation in emergency operations centers, equipping decision makers with data-driven tools to balance speed, safety, and social equity in high-stakes evacuation planning.
Get full access to this article
View all access options for this article.