Optimal operation of regional integrated energy system based on multi-agent deep deterministic policy gradient algorithm

The complex energy coupling and uncertainties of renewable energy and load make the dynamic scheduling of the integrated energy system (IES) very difficult. Therefore, an optimal operation method based on the multi-agent deep deterministic policy gradient algorithm (MADDPG) is proposed. Firstly, the IES model is established, and the optimal scheduling problem is transformed into the Markov decision problem; Then, the action-space, state-space and reward function of agents are designed which control energy conversion equipment and energy storage equipment respectively. Furtherly, a multi-agent framework is established based on the MADDPG; Finally, a scheduling simulation example was carried out. The simulation results indicate that compared with the single agent, the multi-agent framework can improve the stability of training for agents and the ability to explore the optimal solution.