Plug-in electric vehicles (PEVs) grow quickly in recent years. The Electric Power Research Institute (EPRI) expects that the PEV market penetration may be 80% by the year 2050. The governments also make policies to stimulate the PEV sales.

However, there are some drawbacks in an unregulated PEV recharging control algorithm. An unwanted load peak that is caused by the PEV recharging could lead to a disruption, such as a brownout or a blackout in the grid.

The thesis proposes a two-stage dynamic stochastic optimization method in order to minimize recharging cost without raising the peak load. The method is an online scheduling algorithm that bases on an online optimization. Our dynamic algorithm depends on the knowledge of driving behaviors and marketing information to take stochastic factors from different angles into account. The method is robust to several kinds of stochastic parameters, has a low communication requirement, and benefits both users and the power utility. In the paper, the system structure, data models, and mathematical formulation of the proposed method are introduced.