EAGLE: Edge-Aware Graph Learning for Proactive Delivery Delay Prediction in Smart Logistics Networks

Zhiming Xue; Menghao Huo; Yujue Wang

cs.AIApr 6, 2026

EAGLE: Edge-Aware Graph Learning for Proactive Delivery Delay Prediction in Smart Logistics Networks

Authors: Zhiming Xue, Menghao Huo, Yujue Wang

Organizations: College of Engineering, Northeastern University Boston, USA · School of Engineering, Santa Clara University Santa Clara, USA · Department of Chemistry and Chemical Biology, University of New Mexico Albuquerque, USA

Abstract

Modern logistics networks generate rich operational data streams at every warehouse node and transportation lane -- from order timestamps and routing records to shipping manifests -- yet predicting delivery delays remains predominantly reactive. Existing predictive approaches typically treat this problem either as a tabular classification task, ignoring network topology, or as a time-series anomaly detection task, overlooking the spatial dependencies of the supply chain graph. To bridge this gap, we propose a hybrid deep learning framework for proactive supply chain risk management. The proposed method jointly models temporal order-flow dynamics via a lightweight Transformer patch encoder and inter-hub relational dependencies through an Edge-Aware Graph Attention Network (E-GAT), optimized via a multi-task learning objective. Evaluated on the real-world DataCo Smart Supply Chain dataset, our framework achieves consistent improvements over baseline methods, yielding an F1-score of 0.8762 and an AUC-ROC of 0.9773. Across four independent random seeds, the framework exhibits a cross-seed F1 standard deviation of only 0.0089 -- a 3.8 times improvement over the best ablated variant -- achieving the strongest balance of predictive accuracy and training stability among all evaluated models.

Figures & tables

Figure 1. EAGLE three-module architecture pipeline. (1) PatchTST-Lite: Each node’s $T{=}14$ time-step feature sequence is tokenised into $N_{p}{=}2$ non-overlapping patches of length $p{=}7$ , encoded by a 2-layer Transformer (2 attention heads, $d_{\text{model}}{=}64$ ), and mean-pooled to a fixed temporal embedding $\mathbf{h}_{t}\in\mathbb{R}^{64}$ . (2) E-GAT: A 2-layer Edge-Aware GAT (4 heads per layer) aggregates neighbourhood information, conditioning attention coefficients jointly on node embeddings and static edge features $\mathbf{E}_{uv}$ (historical transit statistics, shipping mode distributions). (3) DualHead: Two independent MLP heads produce the SLA-violation probability $\hat{p}_{v}\in[0,1]$ (sigmoid) and expected delay magnitude $\hat{d}_{v}\in\mathbb{R}^{+}$ (Softplus), trained end-to-end via a $0.7{\times}\text{BCE}+0.3{\times}\text{Huber}$ multi-task loss. EAGLE three-module architecture pipeline.

Method	F1 (Macro)	AUC-ROC
XGBoost	$0.6379\pm 0.0074$	$0.7455\pm 0.0020$
Random Forest	$0.6052\pm 0.0018$	$0.7249\pm 0.0032$
LSTM	$0.8095\pm 0.0035$	$0.9679\pm 0.0013$
GAT	$0.6142\pm 0.0789$	$0.6872\pm 0.0224$
EAGLE (Ours)	$\mathbf{0.8762\pm 0.0089}$	$\mathbf{0.9773\pm 0.0037}$

Table 1. Performance comparison on the DataCo test set (mean

\pm

std, 4 seeds). All methods are evaluated at the node-window level with 14-day sliding window snapshots and an identical relative label definition. For tabular models (XGBoost, RF), predicted probabilities at the order level are averaged within each (node, label window) pair.

Figure 2. EAGLE training dynamics across 4 random seeds. (a) Training loss converges smoothly to $\approx 0.186$ by epoch 40. (b) Validation AUC-ROC stabilises above 0.97 after 15 epochs. Stars mark the best-checkpoint epoch per seed. EAGLE training dynamics across 4 random seeds.

† A3 has no regression head; MAE equals the zero-prediction baseline.
Variant	F1 (Macro)	AUC-ROC	MAE (Days)
A1: No Temporal	$0.6851\pm 0.0589$	$0.8605\pm 0.0698$	$0.1801$
A2: No Edge Features	$0.8027\pm 0.0338$	$0.9540\pm 0.0462$	$0.0269$
A3: Single-Task	$0.8230\pm 0.0341$	$0.9738\pm 0.0120$	$0.0731^{\dagger}$
EAGLE (Full)	$\mathbf{0.8762\pm 0.0089}$	$\mathbf{0.9773\pm 0.0037}$	$\mathbf{0.0185}$

Table 2. Ablation Study Results, and MAE (Days) for the regression sub-task.

Figure 3. Supply chain risk graph with per-node risk scores aggregated across all test snapshots. Node color and size encode the normalized risk score (darker red = higher risk); edges represent active shipping lanes. High-risk hubs (Nodes 10, 22, 17) cluster at the network core, while peripheral low-volume nodes maintain consistently low risk. Supply chain risk graph showing per-node normalized risk scores. Node 10 at the network core has the highest risk score of 1.00, shown in red.

Feature	Type	Source	Time Scope	Future Info?	Justification
order_vol	Node	count(orders)	$[t,t+14)$	No	Feature window only
mean_scheduled_transit	Node	mean(scheduled_days)	$[t,t+14)$	No	Known at order placement
std_scheduled_transit	Node	std(scheduled_days)	$[t,t+14)$	No	Known at order placement
mean_discount_rate	Node	mean(discount)	$[t,t+14)$	No	Known at order placement
prev_delay_days	Node	mean(actual_delay)	$[t,t+14)$	No	Past-realised outcomes only
transit_mean	Edge	mean(scheduled_days)	Global static	No	Scheduled (not actual) transit; stable over time

Table 3. Feature audit. Temporal bounds and source for features are recorded to ensure that outcome information does not leak into the features.

Explore similar work

LLM-Grounded Explainable AI for Supply Chain Risk Early Warning via Temporal Graph Attention Networks
Mar 5, 2026Zhiming Xue, Yujue Wang, Menghao HuoSupply Chain Management LLM Reasoning with Graphs
College of Engineering, Northeastern University, Boston, USA · Department of Chemistry and Chemical Biology, University of New Mexico, Albuquerque, USA · School of Engineering, Santa Clara University, Santa Clara, USA
Uncertainty-Aware Delivery Delay Duration Prediction via Multi-Task Deep Learning
Feb 23, 2026Stefan Faulkner, Reza Zandehshahvar, Vahid Eghbal Akhlaghi +3Supply Chain Management Multi-Task Learning
NSF Artificial Intelligence Institute for Advances in Optimization, H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology, Atlanta, GA, USA · Kinaxis
Travel Time Prediction in Supply Chain Management Using Machine Learning
Sep 5, 2026Balaji VenkateswaranSupply Chain Management
Swiss School of Business and Management, Geneva

Mar 5, 2026J/K move · Enter open · S save

LLM-Grounded Explainable AI for Supply Chain Risk Early Warning via Temporal Graph Attention Networks

Zhiming Xue, Yujue Wang, Menghao Huo

College of Engineering, Northeastern University, Boston, USA · Department of Chemistry and Chemical Biology, University of New Mexico, Albuquerque, USA · School of Engineering, Santa Clara University, Santa Clara, USA

Supply Chain Management LLM Reasoning with Graphs GNN Explainability

Open paper AskPDF arXiv

Disruptions at critical logistics nodes pose severe risks to global supply chains, yet existing risk prediction systems typically prioritize forecasting accuracy without providing operationally interpretable early warnings. This paper proposes an evidence-grounded framework that jointly performs supply chain bottleneck prediction and faithful natural-language risk explanation by coupling a Temporal Graph Attention Network (TGAT) with a structured large language model (LLM) reasoning module. Using maritime hubs as a representative case study for global supply chain nodes, daily spatial graphs are constructed from Automatic Identification System (AIS) broadcasts, where inter-node interactions are modeled through attention-based message passing. The TGAT predictor captures spatiotemporal risk dynamics, while model-internal evidence -- including feature z-scores and attention-derived neighbor influence -- is transformed into structured prompts that constrain LLM reasoning to verifiable model outputs. To evaluate explanatory reliability, we introduce a directional-consistency validation protocol that quantitatively measures agreement between generated risk narratives and underlying statistical evidence. Experiments on six months of real-world logistics data demonstrate that the proposed framework outperforms baseline models, achieving a test AUC of 0.761, AP of 0.344, and recall of 0.504 under a strict chronological split while producing early warning explanations with 99.6% directional consistency. Results show that grounding LLM generation in graph-model evidence enables interpretable and auditable risk reporting without sacrificing predictive performance. The framework provides a practical pathway toward operationally deployable explainable AI for supply chain risk early warning and resilience management.

Abstract

Explore similar work

LLM-Grounded Explainable AI for Supply Chain Risk Early Warning via Temporal Graph Attention Networks

Uncertainty-Aware Delivery Delay Duration Prediction via Multi-Task Deep Learning

Travel Time Prediction in Supply Chain Management Using Machine Learning