Achieving Privacy Protection Using Distributed Load Scheduling: A Randomized Approach

With the development of smart grid technology, smart meters have been widely used in various applications, which also incurs security and privacy issues in the meanwhile. With high-precision smart meter data, the residential users' living habits or behavior patterns can be possibly identified for malicious use. To protect the users' privacy, most existing works try to flatten or randomize the aggregated load profile only using rechargeable batteries through centralized control strategies, which may be inefficient and economically costly. In this paper, we propose a general framework to minimize a weighted sum of privacy leakage, electricity bill, and the user experience sacrifice by exploring the capacity of all end-devices of residential users including three categories of shiftable devices, i.e., thermostatically controlled devices, rechargeable batteries, and successive operation devices. By adopting the randomized approach, we propose distributed algorithms to solve the optimization problem efficiently for two scenarios where the user energy is able to feed back to the grid or not. We show that the algorithms can converge to near-optimal solutions with rigorous proofs. Extensive simulations are conducted to demonstrate the effectiveness of proposed algorithms.