Abstract

The efficient and smooth operation of the transportation system is crucial for ensuring the normal functioning of modern society and people’s daily lives. However, the increase in vehicles has led to more frequent and severe traffic congestion due to the limited capacity of existing road networks. Therefore, effectively utilizing the limited traffic capacity of the existing system and reducing the probability of congestion through proper traffic flow management is a critical problem to be solved. In this context, timely and accurate traffic flow forecasting plays an essential role in specifying reasonable traffic control strategies and directly impacting the effectiveness of traffic control measures. However, as a typical large-scale complex network, the movement of traffic on different sections of the road network affects each other. Existing common single-point-based traffic flow prediction methods lack the ability to predict short-term anomalies in traffic flow due to a limited understanding of real-time changes in co-existing traffic flows within the system. Additionally, the scale of the road network directly affects the computational complexity of existing machine learning-based prediction algorithms, resulting in limited scalability. To address these issues, this paper proposes a mixed method that incorporates both historical data and road networks. Our mixed method approach effectively incorporates prior knowledge of the road network into the prediction process, resulting in more accurate predictions and better scalability. Simulation experiments based on a real-world dataset demonstrate a substantial improvement in the prediction accuracy of our method compared to conventional machine learning-based methods. This indicates significant potential benefits, as our method can provide decision-makers with more accurate and timely information for managing traffic flow, leading to improved efficiency, reduced energy consumption, and minimized environmental impact. Therefore, our proposed mixed method has the potential to make a meaningful impact on traffic management and related fields, ultimately contributing to a more sustainable and efficient transportation system.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call