Cluster-Wise Response Aggregation-Based Differentially Private Pool Energy Market Model

This paper¹¹A full financial support is provided in the Core Research Grant by the Science and Engineering Research Board (SERB), India, under IIT Ropar project no. CRG/2018/000084, to carry out this research work. proposes a differential privacy (DP) mechanism for the decentralised pool-based local energy market (PEM) model in the distribution grid. Many traditional PEM models employ alternating direction method of multipliers (ADMM) to reduce the confidential information sharing between participants and energy market operator (EMO). However, an EMO/Aggregator with malicious intent or any cyber-attacker reading the communication between the EMO and participants can still infer the value by looking into their responses over a large time-span. To address the challenge, the proposed DP mechanism completely masks these shared exact information without altering the PEM output from the traditional one. The masking is achieved by three stages, namely 1) clustering, 2) random sharing, and 3) aggregation. The efficacy of the proposal is validated using the Pecan street dataset on MATLAB with Gurobi solver, and a very high degree of DP is observed without any loss of optimality in PEM clearing.