Patient evacuation optimization for health care facilities during hurricanes
The total cost for weather-related disasters in the U.S. increases over time, and hurricanes usually create the most damage. Patient evacuation missions continue to be one of the most prevalent challenges, present in almost every major hurricane event. This research develops a comprehensive modeling and methodological framework for a large-scale patient evacuation problem when an area is faced with a forecasted disaster such as a hurricane. In this work, a hurricane scenario generation scheme using hydrological models for forecasting inland and coastal flooding and a scenario-based stochastic integer program are integrated to make decisions on patient movements, staging area locations and positioning of emergency medical vehicles. The objective is to minimize the total expected cost of evacuation and the setup cost of staging areas. The hurricane scenario generation scheme incorporates uncertainties in the hurricane intensity, direction, forward speed, and tide level. To obtain hurricane scenarios, a hurricane landfall distribution is developed, and stratified sampling is applied to generate potential flooding scenarios. This research also explores different stochastic integer programming formulations of patient evacuation operation, and chooses the most efficient approach for representing the unique characteristics of patient evacuation operation. Finally, to demonstrate the modeling approach, real-world data from the Southeast Texas region is used in numerical experiments. These experiments address exact and approximate methods for finding solutions of the patient evacuation problem for a large network. These results also highlight the importance of operation time limits, the number of available resources, when making evacuation decisions.