XG-DelayNet: An Explainable Graph-Based Machine Learning Framework for Predicting Transportation Delay Propagation in Global Containerized Supply Networks
Abstract
Transportation delay prediction in global supply networks remains challenging because containerized shipments pass through multiple connected actors and locations, including ports, carriers, terminals, customs checkpoints, inland corridors, and transshipment hubs. Although existing machine learning studies in logistics and maritime transportation have improved shipment-level delay prediction, vessel ETA forecasting, and operational risk classification, many models still examine delays as separate prediction events rather than as dynamic disruptions that can spread across interconnected supply-network nodes. than a network-propagated disruption phenomenon. This study develops XG-DelayNet, an explainable graph-based machine learning framework for predicting transportation delay occurrence, delay duration, and delay propagation risk in global containerized supply networks. The framework models ports, terminals, customs points, and inland logistics hubs as nodes, while shipment routes and trade-lane connections are represented as weighted edges. Shipment attributes, carrier schedule reliability, port dwell time, The prediction inputs are weather severity, customs clearance time, transshipment intensity, edge delay history, and network centrality. We compare the proposed temporal graph neural network with Logistic Regression, Random Forest, XG Boost, Light GBM, LSTM and non-graph neural network. To avoid over-estimating performance, we employ a combination of stratified train-test split, recurrent cross-validation, and out-of-sample validation across shipment routes. The results show that XG-Delay Net has the best performance for the prediction of delay occurrences, with an accuracy of 0.901, F1-score of 0.859, and ROC-AUC of 0.946. The model achieves the lowest MAE of 8.72 hours in predicting delay length, indicating a better error reduction compared to the baseline models. The explainability research shows that port dwell time deviation, carrier schedule unreliability, node betweenness centrality, edge delay history and transshipment count are the most prominent sources of delay. In conclusion, the paper proposes an interpretable network-based approach for understanding and forecasting the emergence and transmission of transportation delays across global containerized supply networks
Article Information
Journal |
International Journal of Science, Research and Technology |
|---|---|
Volume (Issue) |
Vol. 7 No. 3 (2024): International Journal of Science, Research and Technology (IJSRAT) |
DOI |
|
Pages |
12117-12135 |
Published |
June 20, 2024 |
| Copyright |
All rights reserved |
Open Access |
This work is licensed under a Creative Commons Attribution 4.0 International License. |
How to Cite |
Gautam Kumar (%2024). XG-DelayNet: An Explainable Graph-Based Machine Learning Framework for Predicting Transportation Delay Propagation in Global Containerized Supply Networks. International Journal of Science, Research and Technology , Vol. 7 No. 3 (2024): International Journal of Science, Research and Technology (IJSRAT) , pp. 12117-12135. https://doi.org/10.15662/IJSRAT.2024.0703005 |
References
2. Baryannis, G., Dani, S., & Antoniou, G. (2019). Predicting supply chain risks using machine learning: The trade-off between performance and interpretability. Future Generation Computer Systems, 101, 993-1004. https://doi.org/10.1016/j.future.2019.07.059
3. Dolgui, A., Ivanov, D., & Sokolov, B. (2018). Ripple effect in the supply chain: An analysis and recent literature. International Journal of Production Research, 56(1-2), 414-430. https://doi.org/10.1080/00207543.2017.1387680
4. El Mekkaoui, S., et al. (2022). Machine learning models for efficient port terminal operations. IFAC-PapersOnLine, 55(10), 1815-1820. https://doi.org/10.1016/j.ifacol.2022.09.656
5. Hamilton, W. L., Ying, R., & Leskovec, J. (2017). Inductive representation learning on large graphs. Advances in Neural Information Processing Systems, 30, 1024-1034.
6. Ivanov, D. (2020). Predicting the impacts of epidemic outbreaks on global supply chains: A simulation-based analysis on the coronavirus outbreak. Transportation Research Part E: Logistics and Transportation Review, 136, 101922. https://doi.org/10.1016/j.tre.2020.101922
7. Kim, Y., Chen, Y. S., & Linderman, K. (2015). Supply network disruption and resilience: A network structural perspective. Journal of Operations Management, 33-34, 43-59. https://doi.org/10.1016/j.jom.2014.10.006
8. Kipf, T. N., & Welling, M. (2017). Semi-supervised classification with graph convolutional networks. International Conference on Learning Representations.
9. Li, Y., Zobel, C. W., Seref, O., & Chatfield, D. (2020). Network characteristics and supply chain resilience under conditions of risk propagation. International Journal of Production Economics, 223, 107529. https://doi.org/10.1016/j.ijpe.2019.107529
10. Li, Y., Chen, K., Collignon, S., & Ivanov, D. (2021). Ripple effect in the supply chain network: Forward and backward disruption propagation, network health and firm vulnerability. European Journal of Operational Research, 291(3), 1117-1131. https://doi.org/10.1016/j.ejor.2020.09.053
11. Lundberg, S. M., & Lee, S. I. (2017). A unified approach to interpreting model predictions. Advances in Neural Information Processing Systems, 30, 4765-4774.
12. Ponomarov, S. Y., & Holcomb, M. C. (2009). Understanding the concept of supply chain resilience. The International Journal of Logistics Management, 20(1), 124-143. https://doi.org/10.1108/09574090910954873
13. Ribeiro, M. T., Singh, S., & Guestrin, C. (2016). Why should I trust you? Explaining the predictions of any classifier. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 1135-1144. https://doi.org/10.1145/2939672.2939778
14. Teece, D. J. (2007). Explicating dynamic capabilities: The nature and microfoundations of enterprise performance. Strategic Management Journal, 28(13), 1319-1350. https://doi.org/10.1002/smj.640
15. Velickovic, P., Cucurull, G., Casanova, A., Romero, A., Lio, P., & Bengio, Y. (2018). Graph attention networks. International Conference on Learning Representations.
16. Wenzel, H., Smit, D., & Sardesai, S. (2019). A literature review on machine learning in supply chain management. Artificial Intelligence and Digital Transformation in Supply Chain Management, 413-441.
17. Wieland, A., & Wallenburg, C. M. (2013). The influence of relational competencies on supply chain resilience. International Journal of Physical Distribution & Logistics Management, 43(4), 300-320. https://doi.org/10.1108/IJPDLM-08-2012-0243