Soft Actor Critic-assisted Receding Horizon Generation Redispatch for Power System Resilience Enhancement against Extreme Weather

Extreme weather-related power outages occur more frequently in the worldwide due to the climate change. In this paper, we focus on the receding horizon control during real-time operation in order to enhance the power system resilience against the impending extreme weather. The mathematic formulation for receding horizon generation redispatch is firstly presented. Considering this optimization problem as a Markov decision process, the soft actor critic (SAC)-based deep reinforcement learning algorithm is proposed to enable the real-time decision-making for power system resilience enhancement against the extreme weather events. Case study is presented to demonstrate that the proposed SAC-based strategy can reduce the loss of load and adapt to the uncertain scenarios under the impact of the extreme weather events.