Plug-in Electric Vehicle charging approaches: Centralized versus decentralized and strategic versus cooperative

Typically, the charging of Plug-in Electric Vehicles (PEVs) can be shifted in time. This flexibility can be used to reduce the cost and/or network impact of charging, and even to provide ancillary services. The focus of this paper is on the use of charging flexibility for cost minimization, either from the perspective of PEVs or from the society's point of view. Charging control approaches can be classified along two dimensions: a) The decisions can be taken locally by the PEVs (decentralized control) or centrally by an agent, typically called aggregator, (centralized control), b) The decision-making agent(s) (PEVs or aggregator) can take decisions that are socially optimal (cooperative), or can behave strategically, maximizing their own profit. We describe approaches that correspond to the related four possible configurations, and which take the uncertainty in driving patterns and the market-related uncertainty into account. We perform simulations with real market data and realistic driving patterns to compare the different approaches. Results show that the benefit from exploiting charging flexibility is substantial, even if the agents behave cooperatively. Moreover, driving uncertainty has a minor impact on costs compared to market-related uncertainty.