Physics-Inspired Spatial-Temporal Graph Neural Networks for Predicting Industrial Chain Resilience
Wang BC(王必成); Wang JP(王军平)
Source PublicationJournal of Machine Learning Research
Industrial chains are playing an increasingly important role in the sustainable development of the national economy. However, the precise prediction of the dynamics of industrial chain  evolution has become a significant challenge in the field of machine learning. Currently, in
the aspect of machine learning solving system dynamics predictionsome successful examples have been achieved in this field, such as PINNs (Physics-Informed Neural Networks). However, machine learning methods, represented by PINNs, in describing and analyzing
the dynamical behavior of industrial chain resilience, is still in its infancy. At its core, there is a lack of a theoretical framework for describing system dynamics. This paper proposes physics-inspired spatial-temporal Graph Neural Networks (PhG-Net), which aiming to describe and predict the evolutionary dynamics of industrial chain resilience. The key idea is
to construct the information diffusion equation of graph neural network and embed it into the loss function of graph neural network as a physical inspiration. The spatial and temporal characteristics and node correlations of the spatial-temporal graph neural network are
used to predict the link and node status in the future, and describe the resilience of the industrial chain from two aspects, link and node status. Attack the network and analyze the resilience of the industry chain from the development trend of the network. Experiments were conducted using real industrial chain datasets, and the results proved that our model can obtain better results and make more accurate and effective predictions of industrial chain resilience, which has certain practical significance for industrial development.
MOST Discipline Catalogue工学::控制科学与工程
Indexed BySCIE
WOS Research AreaNational Key Research and Development Program of China under Grant 2022YFF0903 ; National Natural Science Foundation of China under Grant 92167109300
WOS SubjectComputer Science, Artificial Intelligence
Sub direction classification人工智能基础理论
planning direction of the national heavy laboratory认知机理与类脑学习
Paper associated data
Document Type期刊论文
Corresponding AuthorWang JP(王军平)
AffiliationInstitute of Automation, Chinese Academy of Science
Recommended Citation
GB/T 7714
Wang BC,Wang JP. Physics-Inspired Spatial-Temporal Graph Neural Networks for Predicting Industrial Chain Resilience[J]. Journal of Machine Learning Research,2024,4(23):1-23.
APA Wang BC,&Wang JP.(2024).Physics-Inspired Spatial-Temporal Graph Neural Networks for Predicting Industrial Chain Resilience.Journal of Machine Learning Research,4(23),1-23.
MLA Wang BC,et al."Physics-Inspired Spatial-Temporal Graph Neural Networks for Predicting Industrial Chain Resilience".Journal of Machine Learning Research 4.23(2024):1-23.
Files in This Item: Download All
File Name/Size DocType Version Access License
Physics-Inspired Spa(2342KB)期刊论文作者接受稿开放获取CC BY-NC-SAView Download
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[Wang BC(王必成)]'s Articles
[Wang JP(王军平)]'s Articles
Baidu academic
Similar articles in Baidu academic
[Wang BC(王必成)]'s Articles
[Wang JP(王军平)]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[Wang BC(王必成)]'s Articles
[Wang JP(王军平)]'s Articles
Terms of Use
No data!
Social Bookmark/Share
File name: Physics-Inspired Spatial-Temporal Graph Neural Networks for Predicting Industrial Chain Resilience(1).pdf
Format: Adobe PDF
All comments (0)
No comment.

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.