Multi-Agent Reinforcement Learning-Based Distributed Cooperative Voltage Control

The voltage control of modern power systems faces the challenges of complexity and uncertainty brought by a high proportion of renewable energy access. This paper proposes a distributed voltage cooperative control method based on multi-intelligence reinforcement learning, constructs a multi-level intelligent system model, and designs an improved deep deterministic policy gradient algorithm and a multi-objective hybrid reward function. The method overcomes the limitations of traditional control strategies in complex grid topology and high penetration scenarios, and achieves efficient cooperative control of voltage. Experimental validation shows that the method has superior control performance, strong robustness and adaptability, and provides a new idea for voltage stability control in smart grids.