Disruption Risk Evaluation on Large-scale Production Network with Establishments and Products

2024-10-08

arXiv.org

AI Summary

This study addresses supply chain disruption risk assessment in large-scale production networks. We construct a firm-level, multi-product heterogeneous network covering Japan’s entire manufacturing sector—comprising 184,000 firms and 920,000 supplier–buyer links—by integrating geospatial firm data with fine-grained product flow information. We propose a probabilistic disruption propagation model and conduct regional scenario simulations. Our findings reveal three key contributions: (1) the firm-level network exhibits stronger shock amplification than conventional firm-aggregate networks; (2) incorporating product-specific flows significantly improves risk propagation modeling accuracy, mitigating biases arising from coarse industry-level aggregation assumptions; and (3) while disruptions originating in Tokyo show pronounced regional heterogeneity in firm-aggregate networks, this heterogeneity vanishes in the firm-level network, indicating that granular spatial–product structure homogenizes systemic vulnerability. These results establish a new paradigm for precision risk assessment and evidence-based, targeted policy design in supply chain resilience.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Reasoning under Uncertainty: Relational Probabilistic Models
• Search and Optimization: Distributed Search

Application Category

• Social Networks and Social Media: Social media analysis through the lenses of networks
• Security and Privacy: Large-scale security measurements
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

We constructed an establishment-level production network where each establishment inputs and outputs multiple products, using data that includes the firm-level production network and establishments covering nearly all Japanese entities. The network represents the manufacturing sector with 183,951 establishments across 157,537 firms and 919,982 inter-establishment linkages. A probabilistic model of supply chain disruptions was applied to this network. The key findings are as follows: (1) The establishment-level network exhibits greater shock propagation compared to the firm-level network. (2) Incorporating actual product information leads to a larger impact on propagation compared to using industry-level information. (3) Regional shock simulations reveal that while the firm-level network shows greater shock propagation when the shock originates in Tokyo, no such difference is observed in the establishment-level network.

Problem

• Evaluates disruption risks in large-scale production networks.
• Compares shock propagation at establishment vs. firm levels.
• Assesses regional shock impacts using detailed product data.

Innovation

• Constructed establishment-level production network
• Applied probabilistic supply chain disruption model
• Simulated regional shocks on network propagation