Spatial allocation of electric vehicle charging capacity to facilitate disaster evacuation

Limited public charging infrastructure for electric vehicles (EVs) could prove problematic for disaster evacuations. We formulate a two-stage stochastic program that optimally allocates a budget to install EV charging stations at locations in a road network to maximize expected evacuation performance, subject to uncertain disaster impacts on roads and chargers. Charging investments are made in stage one and an energy-limited traffic routing problem is solved in stage two. We apply the model to a California wildfire case study and explore how several factors affect the optimal charging investments and evacuation performance. We find that initial charge levels are the most critical determinant of evacuation performance; it is difficult to compensate for low initial charge levels, even with a high budget and fast charging speed. When settings are more favorable (higher budget, faster charging, higher initial charge levels), it is optimal to distribute EV chargers more evenly across the road network.