Residential Demand Response Strategy Based on Deep Deterministic Policy Gradient

With the continuous improvement of power system and deepening electricity market reform, trend users’ active participation in distribution is more significant. Demand response has become promising focus smart grid research. Providing reasonable incentive strategies for companies demand customers plays a crucial role maximizing benefits different participants. To meet expectations multiple agents same environment, deep reinforcement learning was adopted. The generative model residential strategy under policies can be trained iteratively through real-time interactions with environmental conditions. In this paper, novel optimization strategy, based on deterministic policy gradient (DDPG) algorithm, proposed. proposed work validated actual consumption data certain area China. results showed that DDPG could optimize policies. addition, overall goal peak load-cutting valley filling achieved, which reflects prospects market.