Traffic Signal Dynamic Optimization Algorithm Based on Deep Reinforcement Learning
A traffic signal dynamic optimization algorithm, the Cross-Attention Mechanism and Dueling Double DQN (CAM-D3QN), which utilizes a novel crisscross attention module to dynamically model spatial dependencies between intersections and incorporates the Dueling Double DQN architecture for robust Q-value estimation.