GeneraLight: Improving Environment Generalization of Traffic Signal Control via Meta Reinforcement Learning

TL;DR

GeneraLight employs meta reinforcement learning and traffic flow generation to enhance traffic signal control models' ability to generalize across diverse real-world traffic environments.

Contribution

The paper introduces a novel meta-RL framework combined with a Wasserstein GAN-based traffic flow generator to improve TSC model generalization.

Findings

GeneraLight outperforms baseline models in diverse traffic scenarios.

The traffic flow generator produces high-quality, varied traffic data.

Meta-learning enhances the adaptability of TSC models to new environments.

Abstract

The heavy traffic congestion problem has always been a concern for modern cities. To alleviate traffic congestion, researchers use reinforcement learning (RL) to develop better traffic signal control (TSC) algorithms in recent years. However, most RL models are trained and tested in the same traffic flow environment, which results in a serious overfitting problem. Since the traffic flow environment in the real world keeps varying, these models can hardly be applied due to the lack of generalization ability. Besides, the limited number of accessible traffic flow data brings extra difficulty in testing the generalization ability of the models. In this paper, we design a novel traffic flow generator based on Wasserstein generative adversarial network to generate sufficient diverse and quality traffic flows and use them to build proper training and testing environments. Then we propose a meta-RL TSC framework GeneraLight to improve the generalization ability of TSC models. GeneraLight boosts the generalization performance by combining the idea of flow clustering and model-agnostic meta-learning. We conduct extensive experiments on multiple real-world datasets to show the superior performance of GeneraLight on generalizing to different traffic flows.