Abstract
Volt-VAR control (VVC) is a critical application in active distributionnetwork management system to reduce network losses and improve voltage profile.To remove dependency on inaccurate and incomplete network models and enhanceresiliency against communication or controller failure, we propose consensusmulti-agent deep reinforcement learning algorithm to solve the VVC problem. TheVVC problem is formulated as a networked multi-agent Markov decision process,which is solved using the maximum entropy reinforcement learning framework anda novel communication-efficient consensus strategy. The proposed algorithmallows individual agents to learn a group control policy using local rewards.Numerical studies on IEEE distribution test feeders show that our proposedalgorithm matches the performance of single-agent reinforcement learningbenchmark. In addition, the proposed algorithm is shown to be communicationefficient and resilient.