Deferrable load scheduling optimization under power price information attacks in smart grid

The public utility can implement real-time pricing (RTP) as a demand response (DR) program in smart grid to shed the power consumption (e.g., by reducing consumption during peak period). In this paper, we consider the optimization of deferrable load scheduling to minimize the power consumption cost. The constrained Markov decision process (CMDP) model is formulated and solved to obtain the optimal scheduling policy. In addition, we consider the case that the data communications system to support RTP (i.e., to broadcast power price information to the deferrable load) can be under attack, which makes the power price information unavailable or falsified. The loss due to such attack can be analyzed using the proposed CMDP model. The results show that the power consumption cost in a scenario with attack is higher than that in a scenario without attack. The analysis will be useful for improving the intrusion detection system (IDS) to defend the attack.