Charging scheduling in a workplace parking lot: Bi-objective optimization approaches through predictive analytics of electric vehicle users' charging behavior

Decarbonization of the transportation sector relies on the widespread adoption of Electric Vehicles (EVs) and appropriate charging strategies. However, uncoordinated EV charging can adversely affect the power grid, and effective scheduling schemes are necessary to mitigate adverse effects. This study aims to develop bi-objective optimization models for EV charging scheduling at a workplace charging station, addressing the EV users' preferences in terms of economic and Quality-of-Service (QoS) dimensions, by minimizing the charging cost considering the participation in Vehicle-to-Grid (V2G) schemes and minimizing the deviation from the desired State-of-Charge (SoC). To address this deviation, two perspectives are considered: minimizing the sum of deviations, embodying a compensatory criterion, and minimizing the worst deviation, a fairness criterion based on a min-max approach. To obtain a representation of the non-dominated solution set corresponding to the scheduling plan for each EV, the Epsilon-constraint method is used. Furthermore, machine learning techniques are employed to predict the charging behavior of EV users, including the desired SoC and charging budget. A sensitivity analysis is also conducted to explore the influence of energy selling prices in V2G mode to accommodate EV users' preferences. The findings indicate that as the difference between the energy buying and selling prices increases, it becomes more challenging to satisfy the desired SoC based on the defined charging budget. Additionally, the model that aims to minimize the charging cost and the worst-case deviation to the desired SoC is more sensitive to changes in energy selling prices, highlighting the impact of price variations in scheduling plans.