Resilient Routing: Risk-Aware Dynamic Routing in Smart Logistics via Spatiotemporal Graph Learning

TL;DR

Last-mile logistics face congestion and demand volatility that degrade reliability under static routing. The authors propose RADR, a framework that combines K-Means-based topology construction, a hybrid ST-GNN for congestion risk forecasting, and a risk-aware dynamic routing objective that inflates edge costs as $W_{dyn}(u,v)=dist(u,v)\,(1+\lambda\cdot Risk_{avg})$. The route is selected by $P^*=\arg\min_{P\in \mathcal{P}_{uv}}\sum_{(i,j)\in P} W_{dyn}(i,j)$ and a path risk score $Risk(P)=\sum_{(i,j)\in P} Risk_{avg}(i,j)$, enabling a trade-off between distance and reliability. On Smart Logistics Dataset 2024, RADR reduces congestion risk exposure by 19.3% with only a 2.1% increase in distance, demonstrating practical resilience gains for supply chains.

Abstract

With the rapid development of the e-commerce industry, the logistics network is experiencing unprecedented pressure. The traditional static routing strategy most time cannot tolerate the traffic congestion and fluctuating retail demand. In this paper, we propose a Risk-Aware Dynamic Routing(RADR) framework which integrates Spatiotemporal Graph Neural Networks (ST-GNN) with combinatorial optimization. We first construct a logistics topology graph by using the discrete GPS data using spatial clustering methods. Subsequently, a hybrid deep learning model combining Graph Convolutional Network (GCN) and Gated Recurrent Unit (GRU) is adopted to extract spatial correlations and temporal dependencies for predicting future congestion risks. These prediction results are then integrated into a dynamic edge weight mechanism to perform path planning. We evaluated the framework on the Smart Logistics Dataset 2024, which contains real-world Internet of Things(IoT) sensor data. The experimental results show that the RADR algorithm significantly enhances the resilience of the supply chain. Particularly in the case study of high congestion scenarios, our method reduces the potential congestion risk exposure by 19.3% while only increasing the transportation distance by 2.1%. This empirical evidence confirms that the proposed data-driven approach can effectively balance delivery efficiency and operational safety.

Figures (4)

• Figure 1: The proposed RADR framework architecture
• Figure 2: Geo-distribution Map
• Figure 3: Model Performance. (Left) Training convergence, (Right) Traffic congestion prediction
• Figure 4: Routing Decision Case Study