Energy price forecasting for optimal managing of electric vehicle fleet

Defining tools and algorithms to support the decision-making process for charging electric vehicles (EVs) is a fundamental theme for the spread of EVs. Utilities can use this approach to incentive or discourage the charge of EVs according to different constraints. In this study, the authors refer the EV clusters or fleets, where there is only one energy buyer for all the clusters. This approach corresponds to an indirect method based on prices to induce behaviours in the management of charging on clusters of EVs. The first actor of the algorithm is an aggregator of EV fleet operators acting as a dealer between the electricity market and consumers. A theoretical game model based on Stackelberg's formulation is proposed to capture the interaction between the fleet operator and the owners/drivers of the EVs. A bi-level optimisation problem arises to represent the game between the agents involved: at the upper level, the aggregator maximises its benefits, while the lower level represents the behaviour of rational drivers as a fleet. The proposed method is applied to actual data obtained observing the behaviour of a car-sharing fleet