This article implements and compares two potential algorithms in urban traffic management: the RAO-3 optimization algorithm and the Q-Learning reinforcement learning algorithm. The research is conducted on the dynamic traffic density optimization system platform, which collects real-time traffic data from IoT devices and processes it through fog computing infrastructure. The implementation of RAO-3 and Q-Learning on this rich dataset can be considered a groundbreaking contribution, helping to identify a more optimal algorithm for traffic flow and routing based on current conditions. The core idea of the research is to manually create a set of sample data while also extracting data from the dynamic traffic density optimization system, then testing this dataset with the RAO-3 and Q-Learning algorithms. The results indicate that Q-Learning outperforms RAO-3 in terms of efficiency and accuracy. This serves as a foundation for future advancements in smart city technology, emphasizing the role of integrating advanced technology in promoting more sustainable, efficient, and safer urban environments.
Le et al. (Thu,) studied this question.