Robust valet charging for electric vehicles with uncertain service time

This paper focuses on the scheduling problem of a new charging service model called valet charging service. In this service model, the service provider sends staff to charge electric vehicles for customers to reduce their range anxiety. When the calls for valet charging service are on a scale, the charging station designation and routing plan for staff will be essential for the service operation. Since these service requirements usually have time window constraints, and the valet charging service time is usually uncertain, the charging station designation will significantly impact the staff's service routing plan, and vice versa. Therefore, in this study, considering the uncertainty of service time, the charging station assignments and routing plans are determined simultaneously. The well-known budgeted uncertainty set is constructed to capture the uncertainty of service time. A novel robust counterpart model is introduced by incorporating dynamic program recursive equations into the deterministic model, which can be solved by the off-the-shelf solvers. In addition, a pulse algorithm integrated branch-and-price algorithm is developed, proven efficient in solving different scale-size problems. Based on a case study in Huangpu District in Shanghai, meaningful managerial insights are provided, which can help the service provider make better implementation decisions.