A stochastic optimization model for staged hospital evacuation during hurricanes. (July 2021)

• Record Type: Journal Article
• Authors: Rambha, Tarun; Nozick, Linda K.; Davidson, Rachel; Yi, Wenqi; Yang, Kun
• Is Part Of: Transportation research. Volume 151(2021)
• Publication Date: 2021-07
• Subjects: Hurricanes -- Hospital evacuation -- Disaster management -- Scenario trees -- Multi-stage stochastic programming; Logistics -- Periodicals; Transportation -- Periodicals; 388.011
• DOI: 10.1016/j.tre.2021.102321
• Languages: English
• ISSNs: 1366-5545

Abstract

Highlights: The problem of staging hospital evacuations during hurricanes is studied. A scenario tree-based stochastic optimization model is used to reduce risk and cost. We determine when and how different types of patients must evacuate to receiving hospitals. Decisions are adaptive to evolving hurricane, roadway traffic, and flood conditions. A hypothetical evacuation of a hospital in North Carolina demonstrates model applications. Abstract: Hurricanes result in large scale evacuations almost every year. Of particular concern and difficulty is the decision of whether or not to evacuate hospitals in these emergencies. During an emergency, a hospital is a source of refuge, and evacuating its patients is often viewed as a last resort since it is difficult to provide quality care while transporting them. At the same time, flooding and loss of power and communications put patients and caregivers at very high risk. Most emergency response plans do not have clear guidelines for evacuating or sheltering-in-place. Hurricanes are particularly complicated because there is often considerable uncertainty surrounding their eventual trajectory and intensity. These factors have contributed to, what is in hindsight, poor decisions that have cost lives. The current paper addresses this problem by developing a stochastic optimization formulation, taking into account evolving conditions and, therefore a hopefully robust collection of future flood, wind, and roadway traffic predictions. The model determines the order in which patients should be evacuated over time based on the evolution of the storm by trading off cost and risk. A holistic case study focused on North Carolina and the evolution of Hurricane Isabel is presented by fusing data and model outputs from different sources. The results highlight the advantages of using a recourse formulation that adapts to new information and illustrates the proposed decision-support model's long-term applications.