2026 Vol.9 Apr.N02

• Title: An Event-Driven Incremental Graph Inference Method for Real-Time Freight Visibility and Exception Response
• Name: Jiawei Li
• Company: Guangzhou Shein International Import and Export Co., Ltd. (SHEIN)，guangzhou，511445，China
• Abstract:

  Real-time freight operations increasingly depend on timely and reliable interpretation of shipment events to support operational visibility, anomaly response, and logistics decision-making. Existing freight event inference methods primarily rely on time-series mechanisms that focus on temporal features of individual nodes, while overlooking the structural dependencies among nodes within freight networks. This limitation reduces inference accuracy and constrains the operational usefulness of event intelligence in complex logistics environments.

   

  To address this problem, this paper proposes an event-driven incremental graph inference method for real-time freight visibility and exception response. A freight association topology graph is first constructed to represent logistics nodes, transport links, and their connectivity relationships. Multi-dimensional node features, including inventory scale, turnaround timeliness, and transport-capacity matching, are then aggregated to form the original node representations. On this basis, an incremental feature propagation mechanism is introduced to perform localized updates under stable graph topology, enabling efficient inference for newly emerging or modified freight events without reconstructing the entire graph.

   

  The proposed method further incorporates perturbation-aware feature correction and inference mapping to classify freight event types and estimate their impact scope. Experimental results on an inter-city freight network dataset show that the proposed method achieves an average event classification overlap of 0.935, outperforming conventional comparison methods and demonstrating stable inference performance. These results indicate that the proposed method can improve the operational interpretability of freight events and provide effective technical support for real-time freight visibility, exception response, and risk-aware logistics coordination.

  
  

• Keyword: real-time logistics intelligence; freight events; incremental graph convolutional networks; inference algorithms;
• DOI: 10.12250/jpciams2026090409
• Citation form: Jiawei Li.An Event-Driven Incremental Graph Inference Method for Real-Time Freight Visibility and Exception Response[J]. Computer Informatization and Mechanical System,2026,Vol.9,pp.

[1] Nareswari A A, Utari D T. ADVANCEMENTS IN ALZHEIMER'S DIAGNOSIS THROUGH MRI USING BAYESIAN CONVOLUTIONAL NEURAL NETWORKS AND VARIATIONAL INFERENCE [J]. BAREKENG: Jurnal Ilmu Matematika dan Terapan, 2024, 18(4):2423–2434.

[2] Wang L, Guo Y, Dong X, et al. Exploring Fine-Grained Sparsity in Convolutional Neural Networks for Efficient Inference [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023, 45(4):4474-4493.

[3] Li J, Yuan P, Zhang J, et al. F2PQNN: a fast and secure two-party inference on quantized convolutional neural networks [J]. Computer Journal, 2025, 68(8):998–1012.

[4] Sander J, Berndt S, Bruhns I, et al. Dash: Accelerating Distributed Private Convolutional Neural Network Inference with Arithmetic Garbled Circuits [J]. IACR Transactions on Cryptographic Hardware and Embedded Systems, 2024, 2025(1):420–449.

[5] Hu Y, Xu X, Duan L, et al. End-Edge Collaborative Inference of Convolutional Fuzzy Neural Networks for Big Data-Driven Internet of Things [J]. IEEE Transactions on Fuzzy Systems, 2025, 33(1):203–217.

[6] Poma Y, Melin P. Prediction Using a Fuzzy Inference System in the Classification Layer of a Convolutional Neural Network Replacing the Softmax Function [J]. Studies in Computational Intelligence, 2024:121–129.

[7] Wang H, Yang S, Shao K, et al. A Privacy-Preserving Convolutional Neural Network Inference Framework for AIoT Applications [J]. Computers, Materials & Continua, 2026, 86(1):1354–1371.

[8] Shang J J, Phipps N, Wey I C, et al. A-DSCNN: Depthwise Separable Convolutional Neural Network Inference Chip Design Using an Approximate Multiplier [J]. Chips, 2023, 2(3):159-172.

[9] Ren W, Li W, Hong Y, et al. GC-PTransE: A Multi-Step Attack Inference Method Based on Graph Convolutional Neural Networks and Translation Embedding [J]. Knowledge and Information Systems, 2025, 67(6):5215–5245.

[10] Zhao Y, Wang K, Louri A. HS-GCN: A High-Performance, Sustainable, and Scalable Chiplet-Based Accelerator for Graph Convolutional Network Inference [J]. IEEE Transactions on Sustainable Computing, 2025, 10(5):1019-1030.

[11] Adeyemo A. A., Sanderson J. J., Odetola T. A., et al. StAIn: Stealthy Avenues of Attacks on Horizontally Collaborated Convolutional Neural Network Inference and Their Mitigation [J]. IEEE Access, 2023, 11:10520–10534.

[12] Liu X, Pu H, Chen Z, et al. EvolGCN: A Co-Evolutionary Graph Convolutional Network Model for Dynamically Spatio-Temporal Anomaly Event Inference [J]. IEEE Transactions on Dependable and Secure Computing, 2025, 22(5):4503–4515.

[13] Lin Z H, Li D. Semantics-guided Adaptive Topology Inference Graph Convolutional Networks for Skeleton-based Action Recognition [J]. Journal of Guangdong University of Technology, 2023, 40(4):45-52.

[14] Kress F, Sidorenko V, Schmidt, Patrick, Hoefer, Julian, Hotfilter, Tim, Walter, Iris, Harbaum, Tanja, Becker, Jurgen. CNNParted: An open source framework for efficient Convolutional Neural Network inference partitioning in embedded systems [J].Computer Networks, 2023, 229(6):109759.1–109759.12.