This paper presents a method for predicting delivery delays and optimizing routes in logistics systems using Graph Neural Networks (GNNs). Modern logistics networks face numerous challenges due to unpredictable delays caused by dynamic traffic conditions, weather events, vehicle malfunctions, and other external factors. Traditional machine learning methods, such as regression models or decision trees, often prove inadequate in modeling such complex spatiotemporal dependencies inherent in logistical environments.
Unlike conventional approaches, Graph Neural Networks are capable of efficiently processing data with an inherent graph structure, such as logistics networks where nodes represent warehouses or delivery points and edges correspond to transportation links. GNNs aggregate information from neighboring nodes and edges, forming vector representations that capture both local and global network characteristics. This enables the system to predict delays at specific segments of a delivery route and respond in advance to emerging risks within the supply chain.
The methodology proposed in this paper combines GNN-based delay prediction with a dynamic route optimization algorithm that adjusts delivery routes in real-time according to updated predictions. This approach is demonstrated on a synthetic example, evaluating the method’s effectiveness using metrics such as average travel time and the probability of on-time delivery. The results highlight the advantages of using graph-based models over standard techniques that fail to consider the network topology.
Additionally, the paper reviews current research in the domain of spatiotemporal Graph Neural Networks (STGNNs), which have demonstrated high accuracy in tasks such as delay forecasting, inventory management, and transportation modeling. The use of GNNs is projected to significantly enhance the efficiency of logistics processes, particularly in complex, highly dynamic networks. As such, the proposed approach represents a valuable contribution to the advancement of next-generation intelligent logistics systems.
- Jiang, W., Han, H., Zhang, Y., Wang, J., He, M., Gu, W., Mu, J., & Cheng, X. (2024). Graph Neural Networks for Routing Optimization: Challenges and Opportunities. Sustainability, 16(21), 9239–9239. https://doi.org/ 10.3390/su16219239
- Liu, T. & Meidani, H. (2023). Graph Neural Networks for Travel Distance Estimation and Route Recommendation Under Probabilistic Hazards. Arxiv.org. https://arxiv.org/html/2501.09803v1
- Long, Q., Fang, Z., Fang, C., Chong, C., Wang, P. & Zhou, Y. (2024). Unveiling Delay Effects in Traffic Forecasting: A Perspective from Spatial-Temporal Delay Differential Equations. Arxiv.org. https:// arxiv.org/html/2402.01231v2
- Narula, A. & Lin, Y. (2024). Recurrent Neural Network for Predicting Sequential Supply Chain Delays. https://ctl.mit.edu/sites/ctl.mit.edu/files/theses/Recurrent%20Neural%20... 0Sequential%20Supply%20Chain%20Delays.pdf
- Shahbazian, R., Pugliese, L., Guerriero, F., & Macrina, G. (2024). Integrating Machine Learning Into Vehicle Routing Problem: Methods and Applications. IEEE Access, 12, 93087–93115. https://doi.org/10. 1109/access.2024.3422479
- Ülkü Küp, B., Küp, E. T., Koçak, G., Yucekaya, A. D., & Hekimoğlu, M. (2025). Real-Time Prediction of Delivery Delay in Supply Chains using Machine Learning Approaches. https://doi.org/10.2139/ ssrn.5062672
- Wang, Y., & Liang, X. (2025). Application of Reinforcement Learning Methods Combining Graph Neural Networks and Self-Attention Mechanisms in Supply Chain Route Optimization. Sensors, 25(3), 955. https://doi.org/10.3390/s25030955
- Wasi, A. T., Islam, S., & Akib, A. R. (2024). Graph Neural Networks in Supply Chain Analytics and Optimization: Concepts, Perspectives, Dataset and Benchmarks. Arxiv.org. https://arxiv.org/html/2411. 08550v1
- Wen, H., Lin, Y., Mao, X., Wu, F., Zhao, Y., Wang, H., Zheng, J., Wu, L., Hu, H., & Wan, H. (2022).
- Graph2Route: A Dynamic Spatial-Temporal Graph Neural Network for Pick-up and Delivery Route Prediction. Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. https://doi.org/10.1145/3534678.3539084
- Wu, Z., Pan, S., Chen, F., Long, G., Zhang, C., & Yu, P. S. (2020). A Comprehensive Survey on Graph Neural Networks. IEEE Transactions on Neural Networks and Learning Systems, 32(1), 4–24. https://doi.org/10. 1109/TNNLS.2020.2978386
- Yan, L., & Xu, Y. (2024). XGBoost-Enhanced Graph Neural Networks: A New Architecture for Heterogeneous Tabular Data. Applied Sciences, 14(13), 5826–5826. https://doi.org/10.3390/app14135826