X-Light: Cross-City Traffic Signal Control Using Transformer on Transformer as Meta Multi-Agent Reinforcement Learner
This addresses the challenge of cross-city traffic management for urban planners and AI researchers, offering a novel transferable solution.
The paper tackles the problem of transferring multi-agent traffic signal control algorithms across diverse cities by proposing X-Light, a Transformer on Transformer model for meta multi-agent reinforcement learning, which achieves an average improvement of +7.91% and up to +16.3% in unseen scenarios compared to baselines.
The effectiveness of traffic light control has been significantly improved by current reinforcement learning-based approaches via better cooperation among multiple traffic lights. However, a persisting issue remains: how to obtain a multi-agent traffic signal control algorithm with remarkable transferability across diverse cities? In this paper, we propose a Transformer on Transformer (TonT) model for cross-city meta multi-agent traffic signal control, named as X-Light: We input the full Markov Decision Process trajectories, and the Lower Transformer aggregates the states, actions, rewards among the target intersection and its neighbors within a city, and the Upper Transformer learns the general decision trajectories across different cities. This dual-level approach bolsters the model's robust generalization and transferability. Notably, when directly transferring to unseen scenarios, ours surpasses all baseline methods with +7.91% on average, and even +16.3% in some cases, yielding the best results.