Recent research has shown that some network traffic flow patterns are similar across multiple traffic regions. Identifying and transferring these domain-invariant features can significantly boost model accuracy and robustness, providing new insights into dealing with modeling issues like traffic data insufficiency and dataset shift. However, how to acquire transferable network traffic flow patterns from multiple traffic regions and adapt such knowledge to downstream prediction tasks of target regions remains challenging. To realize domain-invariant traffic flow pattern transfer and provide more robust prediction under insufficient data conditions, we propose a macroscopic fundamental diagram (MFD) guided transfer learning method, namely physics-guided multi-source domain adversarial network (PG-MDAN). First, an MFD similarity measure is proposed to determine what traffic flow patterns are transferable and to what extent they can be transferred. PG-MDAN embeds this physics-informed transferability measure in domain adversarial pre-training for better adaptation ability. Numerical experiments based on two real-world urban network traffic datasets show that PG-MDAN can successfully transfer recurrent and non-recurrent network traffic flow patterns from multiple regions to provide more robust and responsive prediction performance. Finally, extensive sensitivity analysis is conducted, and the results validate that applying such physical regularization can effectively avoid negative transfer and provide a flexible tool to initiate traffic flow pattern transfer in practice.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Physics-Guided Multi-Source Transfer Learning for Network-Scale Traffic Flow Prediction


    Beteiligte:
    Li, Junyi (Autor:in) / Liao, Chenlei (Autor:in) / Hu, Simon (Autor:in) / Chen, Xiqun (Autor:in) / Lee, Der-Horng (Autor:in)

    Erschienen in:

    Erscheinungsdatum :

    01.11.2024


    Format / Umfang :

    4692180 byte




    Medientyp :

    Aufsatz (Zeitschrift)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    MS-Net: Multi-Source Spatio-Temporal Network for Traffic Flow Prediction

    Fang, Shen / Prinet, Veronique / Chang, Jianlong et al. | IEEE | 2022


    Method for traffic flow prediction by using space-time network guided by curriculum learning

    CHEN XINLONG / FU YANGGENG / CHEN SHOUMING | Europäisches Patentamt | 2025

    Freier Zugriff

    Multi-scale traffic flow prediction method based on graph convolutional neural network

    ZHANG MEIYUE / WANG SENZHANG / MIAO HAO et al. | Europäisches Patentamt | 2021

    Freier Zugriff

    Multi-source Adaptive Fusion Spatiotemporal Network for Traffic Emission Prediction

    Zhang, Guoan / Cao, Yang / Pei, Lihong et al. | Springer Verlag | 2025


    Distributed Multi-Intersection Traffic Flow Prediction using Deep Learning

    Moumen Idriss / Mahdaoui Rabie / Raji Fatima Zahra et al. | DOAJ | 2024

    Freier Zugriff