Optimal selection of depots to facilitate ridesharing during a disaster

Evacuation orders are put in place before or after a disaster to save lives. Despite mandatory evacuation orders due to an imminent disaster, many individuals do not or are not able to evacuate because they do not have access to transportation. We present a network-based optimization model to select a subset of depots, such as public schools, to activate during an evacuation process. These depots facilitate ridesharing by allowing for evacuees without a means of transportation to be matched to volunteers evacuating with a vehicle and at least one empty seat in that vehicle. We present results from a computational study based on data collected during Hurricane Florence. Using the case study, we demonstrate that solutions can be found quickly and that a relatively small number of depots can facilitate a substantial number of matchings.