Fig 1.
Structure of the multi-agent reinforcement learning system integrated with a PNN predictor.
Fig 2.
Probabilistic neural network structure diagram.
Fig 3.
Schematic Diagram of Hierarchical Coordinated Scheduling for Reactive Power and Voltage in Cross-Regional Power Grids Driven by Multi-Agent Reinforcement Learning.
Table 1.
Indicator state classification.
Fig 4.
IEEE 33-node power distribution network topology.
Fig 5.
Multi-agent reinforcement learning grid state prediction results.
Fig 6.
Changes in node variables before and after applying the proposed method.
Table 2.
Statistical Comparison of Performance Metrics (Mean ± Standard Deviation, 30 Runs).
Fig 7.
Comparison of Scheduling Efficiency and Voltage Qualification Rate for the Four Algorithms (Mean ± Standard Deviation over 30 Runs).
Fig 8.
Cumulative Reward Convergence Curve During the Training of the MARL-PC Method.
Table 3.
Paired t-test Results (p-values) for MARL-PC vs. Comparison Methods.