Enhancing Urban Traffic Management in Taipei: A Reinforcement Learning Approach

the study explores the application of reinforcement learning (RL) algorithms and capabilities of the Simulation of Urban Mobility (SUMO) solution to enhance urban traffic management in Taipei. Focusing on two major intersections this research employs Q-learning, a model-free RL algorithm, to optimize traffic signal timings based on real-time transport conditions. The methodology encompasses the vehicles real data collection, as well as traffic light phases, and simulation within the SUMO framework to model urban traffic scenarios. The findings reveal significant improvements in traffic throughput and reductions in trip durations during both peak and non-peak hours, demonstrating the potential of RL algorithms to enhance traffic flow efficiency. The study highlights the algorithm's effectiveness in reducing CO2 emissions, contributing to environmental sustainability goals. The results of the project underscore the importance of adopting advanced computational models in urban traffic management, offering insights into the development of smarter and more sustainable transportation systems.