Related Experiment Videos
Graph-augmented transformer networks and explainable AI for economic impact forecasting in disrupted supply chains
1School of Intelligent Transportation, Luoyang Normal University, Luoyang, 471934, Henan, China. ZhenZhao90@outlook.com.
Scientific Reports
|July 19, 2026
Summary
This study introduces a hybrid AI model for predicting supply chain disruption economic impacts with 3.7% MAPE accuracy. It offers a dynamic resilience score and actionable insights for policymakers to enhance stability.
Area of Science:
- Artificial Intelligence
- Supply Chain Management
- Economic Modeling
Background:
- Global supply chains face increasing disruptions.
- Accurate prediction of economic consequences is crucial for risk management.
- Existing models struggle to capture complex, cascading effects.
Purpose of the Study:
- To develop a hybrid AI model integrating Graph Neural Networks and Transformer models.
- To predict the economic consequences of supply chain disruptions with high accuracy.
- To provide actionable insights for policymakers and decision-makers.
Main Methods:
- Utilized multi-dimensional data from 137 companies across 23 countries.
- Developed a hybrid AI framework combining Graph Neural Networks and Transformer models.
- Implemented a dynamic resilience scoring system and advanced explainability techniques.
Main Results:
- Achieved unprecedented prediction accuracy (MAPE: 3.7%) for economic consequences.
- Demonstrated improved long-term impact prediction by 27.3% compared to traditional models.
- The resilience scoring system achieved 91.4% accuracy.
Conclusions:
- The hybrid AI model offers a robust tool for understanding and mitigating supply chain risks.
- The framework provides a theoretical link between resilience and economic outcomes.
- Results are relevant for strategic and operational decision-making in global supply chains.