A Consensus-Based Distributed Primal-Dual Perturbed Subgradient Algorithm for DC OPF

In this paper, an consensus-based distributed primal-dual perturbed subgradient algorithm is proposed for the DC Optimal Power Flow (OPF) problem. The algorithm is based on a double layer multi-agent structure, in which each generator bus and load bus in electric power grid is viewed as bus agent and connects with the grid by a network agent. In particular, network agents employ the average consensus method to estimate the global variables which are necessary for bus agents to update their generation using a local primal-dual perturbed subgradient method. The proposed approach is fully distributed and realizes the privacy protection. The employment of primal-dual perturbation method ensuring the convergence of the algorithm. Simulation results demonstrate the effectiveness of the proposed distributed algorithm.