An ADMM Based Distributed Finite-Time Algorithm for Economic Dispatch Problems.

In this paper, we present a distributed algorithm based on an alternating direction method of multipliers (ADMM), which is applied to solve economic dispatch problems (EDPs). First, with the help of two indicator functions, an EDP is transformed to an equivalent optimization problem with only equality constraint and thus can be dealt with ADMM. Second, a centralized algorithm is proposed to solve the transformed EDP, and furthermore, a distributed algorithm is designed with the help of finite-time average-consensus control strategy. Compared with the existing algorithms for EDP, the distributed algorithm can solve the economic dispatch problem on directed graphs. Moreover, the proposed algorithms can ensure that the generator constraints are satisfied during the whole computation process. Finally, some simulation results are also provided to demonstrate the effectiveness of the proposed algorithms.