EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Big Data Driven Decision Framework for Predicting Freight Delay Patterns in Multimodal Logistics Networks

Yanqing Li

Data and Metadata · 19.03.2026 · 5 · 1424

DOI: 10.56294/dm20261424 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Data and Metadata, online erschienen am 19.03.2026. Verfasst von Yanqing Li. Die Arbeit nennt 26 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Introduction: rapid expansion of multimodal logistics networks connecting road, rail, sea, and air transportation has complicated freight movement and delays. Modern logistics systems fail to forecast freight delays owing to unforeseen interruptions, different data sources, and changing operating circumstances.Method: this research presents a big data-driven decision framework for forecasting freight delay patterns in multimodal logistics networks (BD-FDPM). The framework incorporates historical and real-time logistics data, including transport routes, traffic situations, weather reports, port and terminal operations, and intermodal transfers. The suggested system uses an advanced data collection and features fusion methods to handle data that is heterogeneous, contains missing values, and varies over time across different modes of transportation. Machine learning (ML) and deep learning (DL) models are combined to find both short-term and long-term delay dependencies. Graph-based analytics is subsequently employed to describe the topology of an intermodal network and how delays spread from one node or connection to another. A decision-support layer turns predictions into useful suggestions for changing routes, switching modes, and changing timings.Results: using multimodal freight datasets for experimental evaluation shows that the proposed framework achieves significantly more accurate and reliable delay prediction than existing statistical and single-mode prediction methods. The proposed BD-FDPM has the smallest predictive errors (MAE: 5.2, RMSE: 7.4, and MAPE: 7.1%), indicating much higher performance compared to all baseline models. It is very reliable and efficient, with prediction reliability of 94.8, delay detection accuracy of 93.6, and congestion hotspot detection accuracy of 92.2. Furthermore, BD-FDPM can compute faster at 1.6s/epoch, and the greatest improvement in the operation is 36.7, making it the most accurate and efficient.Conclusions: the framework lets logistics companies and authorities take steps to avoid delays, improve service reliability, and make multimodal transportation more efficient.

Zitieren

APA

Li, Y. (2026). Big Data Driven Decision Framework for Predicting Freight Delay Patterns in Multimodal Logistics Networks. Data and Metadata, 5, 1424. https://doi.org/10.56294/dm20261424

Vancouver

Li Y. Big Data Driven Decision Framework for Predicting Freight Delay Patterns in Multimodal Logistics Networks. Data and Metadata. 2026;5:1424. doi:10.56294/dm20261424

BibTeX
@article{Li2026,
  author = {Li, Yanqing},
  title = {Big Data Driven Decision Framework for Predicting Freight Delay Patterns in Multimodal Logistics Networks},
  journal = {Data and Metadata},
  year = {2026},
  volume = {5},
  pages = {1424},
  doi = {10.56294/dm20261424}
}
Kostenlos starten →