Multi-Scenario Robust Stochastic Optimization Based Approach for Scheduling of Mobile Charging Stations

Currently, mobile charging stations (MCSs) have emerged as a complement of fixed charging stations. An idle MCS could receive few charging requests due to the lack of neighboring electric vehicles to be charged (EVCs), and thus its charging profit could be affected. To decrease the scheduling cost and increase the charging profit of MCSs, idle MCSs should proactively move to the regions with high charging demand and with low charging supply. However, the future charging demand of EVCs is uncertain and related to some external factors. To this end, we introduce the robust stochastic optimization (RSO) model to formulate the future charging demand as an uncertain variable, and then a scheduling approach based on the multi-scenario robust stochastic optimization (SA-MSRSO) is proposed. In SA-MSRSO, the probability distributions of the uncertain variable across a range of potential scenarios are approximatively obtained by analyzing the historical charging demand, and a multivariate regression tree (MRT) is applied for solving the model. Extensive simulations and comparisons demonstrate the performance superiority of our proposed SA-MSRSO, i.e., the charging profit of MCSs can be significantly increased, and the proportion of successfully charged EVCs can be effectively enhanced.