Multi-Task Reinforcement Learning for Distribution System Voltage Control With Topology Changes
Yansong Pei, Junbo Zhao, Yiyun Yao, Fei Ding
Abstract
This letter proposes a multi-task deep reinforcement learning (DRL) approach for distribution system voltage regulation considering topology changes via PV smart inverter control. The key idea is to encode the topology as an additional state for the DRL and leverage the multi-task learning scheme for joint learning of all task control policies. Unlike other DRL-based methods, our approach is robust to different topologies. Comparison results on the modified IEEE 123-node system demonstrate the enhanced robustness of the proposed method.
Topics & Concepts
Reinforcement learningNetwork topologyComputer scienceRobustness (evolution)Topology (electrical circuits)Leverage (statistics)Artificial intelligenceEngineeringElectrical engineeringComputer networkChemistryBiochemistryGeneOptimal Power Flow DistributionSmart Grid Energy ManagementMicrogrid Control and Optimization