Model Specification and Data Aggregation for Emergency Services Facility Location
Disaster response planning has become an area of much recent focus due to many high-profile Black Swan events occurring in the past few years. Deterministic spatial location models are often used in such plans, and these have been found to be severely impacted by data aggregation. In this paper, we use a computational approach to explore the relative impact of aggregation errors, model choice errors, and their interaction, for an ambulance station facility location model. We compare two model choices —probabilistic and deterministic—using a year of call data from the Edmonton (Canada) Emergency Medical Service. We thereby demonstrate that model choice error dominates aggregation error, which implies that aggregation error is relatively unimportant if one uses the more realistic probabilistic model. The results are robust to perturbations in the way travel distances are measured and in the spatial demand distribution.