Addressing Economic Dispatch Problem in a Smart Grid: A Privacy-Preserving and Distributed Optimization Approach

The smart grid is capable of better scheduling power generation and consumption in an economical manner. The economic dispatch problem (EDP) in a smart gird is concerned with finding how much power each generator should generate for the given demand while minimizing the total operational costs. Currently, there are many effective algorithms for EDP, but most of them do not protect some sensitive information such as the final output of each generator. In this paper, we propose a novel privacy-preserving distributed consensus (PP-DC) algorithm to address EDP, where only row stochastic matrix is required. A customized noise generation mechanism is designed in PP-DC to protect the privacy of the sensitive information. We theoretically prove PP-DC achieves ( \(\varTheta , \varSigma \) )-privacy. Additionally, PP-DC introduces an expression with an auxiliary variable to deal with directed network topology, which includes plenty application scenarios. Furthermore, the non-coordinated constant step-size strategy is adopted in PP-DC. The proposed PP-DC achieves privacy-preserving, adjustable convergence error, and fully distributed simultaneously. Finally, the correctness and effectiveness of PP-DC are confirmed by the experiments.