Predictive supply chain disruption control framework using casual network-based multi-stream deep learning

초록

The dynamic changes in natural disasters, such as the sudden occurrence of typhoons and heavy rains, affect transportation and industrial networks and have a cascading impact on the related supply chains. The predictability of traffic conditions and related supply chains is essential for cost reduction, efficient resource allocation, and customer satisfaction. In this study, we propose a deep learning framework that examines the impact of sudden accidental factors on traffic networks and supply chain management (SCM) networks using causal networks. The framework proposed in this study integrates dynamic and real-time traffic data as well as SCM-related data, which are often altered by natural disasters, into a graph-based platform and organizes them into a novel multi-stream deep learning architecture. The multi-stream deep learning framework based on the proposed dynamic graph effectively models the impact of causal network changes on the routes and SCM networks, and uses it for supply chain control. The performance superiority of the proposed framework is empirically demonstrated by comparing it with various deep learning-based SCM prediction models.

키워드