Energy System Fault Reconstruction Control Strategy Based on Multi-agent Deep Reinforcement Learning

In order to achieve intelligent management of faults in the energy system of stratospheric airships, this paper proposes an energy system fault reconstruction control strategy based on multi-agent deep reinforcement learning. For the architecture of distributed energy cabins, an agent is equipped for each energy cabin. Multiple agents share the same dueling double deep Qlearning (D3QN) strategy to accelerate the training process of the agents. A simulation environment was set up in the Gym of Open AI, and the results showed that the method proposed in this paper can effectively achieve rapid reconstruction of energy systems after single point and multi-point simultaneous failures.