Reinforcement learning for traffic signal control in large-scale transportation networks: a systematic literature review

Multi-Intersection Traffic Signal Control Based on Multi-agent Reinforcement Learning: A Cooperative Approach

A novel cooperative MARL-based approach for adaptive traffic signal control in multi-intersection networks that significantly outperforms existing methods in relation to average pheromone intensity, average noise emission, and average waiting time is proposed.

HiLLTS: Zero-Shot Hierarchical LLM-Guided Traffic Signal Control for Sustainable Transportation

HiLLTS, an LLM-guided traffic signal control framework that employs a hierarchical three-layer architecture consisting of a central coordination agent, a district layer and multiple cluster-level intersection agents, is proposed.

Adaptive traffic control systems in intelligent transportation systems: a systematic literature review

Traffic congestion continues to pose a major challenge, particularly in urban areas where infrastructure capacity and data availability are limited. An adaptive traffic control system (ATCS), part of intelligent transportation systems (ITS), can improve road network performance by dynamically adjusting signal timing in response to real-time traffic conditions. However, research on ATCS remains scattered and diverse, making it difficult to obtain a comprehensive picture of its trends, methods, challenges, and effectiveness. This study seeks to deliver a comprehensive overview of ATCS through a systematic literature review (SLR) involving 77 full-text articles using the preferred reporting items for systematic reviews and meta-analyses (PRISMA) methodology. The analysis focuses on four main aspects: research trends, implementation barriers, methods and technologies used, and performance evaluation metrics. The study’s findings indicate a significant increase in publications after 2019, with simulation-based studies dominating and major contributions coming from China and the United States. Methodological advancements have gradually transitioned from traditional rule-based approaches toward more advanced techniques, including reinforcement learning (RL), deep RL, multi-agent RL, and hybrid approaches. Various enabling technologies, such as edge computing, computer vision, floating-car data (FCD), sensor fusion, digital twins, and vehicle-to-everything (V2X) communication, are increasingly being integrated. However, real-world implementation still faces various challenges, including data limitations, low connected vehicle (CV) penetration, gaps between simulation and field conditions, and security and infrastructure readiness issues. Performance evaluation is also evolving into a multidimensional process encompassing mobility, safety, the environment, and system sustainability. Overall, there is no single universally optimal method; the performance of ATCS is strongly influenced by the nature of the available data and the level of system readiness. Therefore, future research should focus on improving field validation, standardizing evaluation methods, developing solutions that adapt to local conditions, and strengthening system security and reliability to support real-world implementation.

Deep reinforcement learning-based traffic signal control in multi-intersection environments: a comparative study of DQN variants

The findings demonstrate the potential of DRL-based traffic signal control in controlled simulation conditions and highlight that algorithm performance is strongly influenced by traffic policy design and environmental complexity.

Optimizing Traffic Signal Control Using Reinforcement Learning Methods: Hybrid Approach

Urban traffic congestion remains a major challenge for modern cities, requiring intelligent traffic signal control (TSC) strategies capable of adapting to dynamic traffic conditions. This paper proposes a hybrid reinforcement learning approach for traffic signal control that combines the complementary learning mechanisms of Q-learning, SARSA, and Monte Carlo algorithms to improve both learning efficiency and control performance. The proposed approach is implemented and evaluated using the Simulation of Urban MObility (SUMO) simulator on a realistic road network corresponding to the “Route de Sefrou” in Fez, Morocco. The traffic signal controller is trained through continuous interaction with the simulated environment and compared with the three individual reinforcement learning algorithms under identical experimental conditions. The experimental results demonstrate that the proposed hybrid approach provides more efficient traffic management, faster convergence, and greater learning stability than the individual algorithms. These findings demonstrate the potential of hybrid reinforcement learning as an effective solution for adaptive traffic signal control in realistic urban environments.

A Fusion Approach of Reinforcement Learning and Traffic Engineering in Multiple Traffic Signal Control