Whether to charge an electric vehicle or not? A near-optimal online approach

The electric vehicle (EV), a promising technique to reduce transportation emissions, is one of the most important household appliances for demand response. Under the real-time pricing environment in smart grid, EV owners are faced with the EV charging scheduling problem to minimize electricity cost. Existing works focused on offline solutions or Markov decision processes which require full or statistical priori knowledge of future real-time prices (RTPs). This may not be practical as it is difficult to predict future RTPs especially for long horizons. In this paper, we propose a near-optimal online algorithm via primal-dual approach, requiring very little priori knowledge, i.e., the upper bound of future RTPs. The competitive ratio of the online algorithm is derived. It is demonstrated with real price data from Ameren Corporation that our proposed online algorithm can result in considerable economic savings, compared with existing schemes which only consider RTPs.