Emergency shelter location–allocation analysis with time–varying demand

Abstract

The reduction in the overall evacuation time by increasing the number of shelters is desirable. However, policymakers often face resource constraints that limit their ability to open additional shelters. Using the emergency shelter location-allocation model, this study empirically identifies which shelters should be opened given the demand for evacuation varies based on time of day (time-varying demand for each hour) in Higashihiroshima city, Japan. To achieve this, a framework was developed to estimate the time-varying evacuation demand using secondary data sources, followed by an analysis using a location-allocation optimization model. The results show that the dynamic change in evacuation demand over time significantly affects the number and location of shelters to be opened. These findings highlight the importance of time-dependent estimation in disaster response management.