Bidirectional energy supply logistics using uncrewed electric aerial and ground vehicles: A two-echelon location-routing problem with resource-constrained demand allocation and time windows

• Two-echelon bidirectional energy supply logistics model with UGVs and UAVs is formulated. • A two-stage method solves the location-routing and demand allocation problems in each stage. • Sensitivity analysis is conducted under various budget constraints. • The model is applied to a real world post-wildfire case study, improving energy supply and resilience. In the aftermath of hazards, energy supply to demand nodes is constrained due to power outages caused by damaged infrastructure. Electric uncrewed ground vehicles (UGVs) and aerial vehicles (UAVs) can play a crucial role in providing backup power, discharging energy directly to meet demand, and recharging en-route if necessary. However, charging infrastructure that replenishes backup power has limited capacity and can be inoperable due to disruptions of the main power system. Since certain energy demands are urgent in post-disaster contexts (i.e., healthcare loads), UGVs and UAVs prioritize these critical needs. As a result, UGVs and UAVs may meet only portion of the energy needs due to constrained resources. This research aims to design a bidirectional energy supply logistics model, as a two-echelon electric vehicle location-routing problem with resource-constrained demand allocation and time windows, employing UGVs and UAVs. The first echelon involves deploying UGVs and UAVs from a depot to satellite charging locations. In the second echelon, these vehicles travel from satellites to serve local energy demand, either fully or partially within the constrained resource budget discharging their batteries and recharging en-route as needed. We formulate this model and propose a metaheuristic solution which consists of a two-stage approach based on the adaptive large neighborhood search. Numerical experiments on benchmark instances were conducted to evaluate the novel heuristic’s performance compared to a commercial solver. Sensitivity analysis was carried out to examine the impact of UAV and UGV batteries capacity, budget for charging infrastructure, and the amount of energy resources. We applied our model to a real-world case of post-wildfire humanitarian aid in California’s counties aiming to supply energy during large-scale power loss in the region.