Dynamic Scheduling Fusion Model for Railway Hazardous Chemical Transportation Emergency Supplies Based on DBSCAN–Bayesian Network

Hao Yin, Minbo Zhang (), Chen Lei, Kejiang Lei, Tianyu Li and Yuhao Jia
Additional contact information
Hao Yin: School of Resource and Safety Engineering, Wuhan Institute of Technology, Wuhan 430074, China
Minbo Zhang: School of Resource and Safety Engineering, Wuhan Institute of Technology, Wuhan 430074, China
Chen Lei: Graduate School, University of International Business and Economics, Beijing 100029, China
Kejiang Lei: School of Resource and Safety Engineering, Wuhan Institute of Technology, Wuhan 430074, China
Tianyu Li: China Railway Zhengzhou Bureau Group Co., Ltd. Institute of Science and Technology, Zhengzhou 450052, China
Yuhao Jia: School of Resource and Safety Engineering, Wuhan Institute of Technology, Wuhan 430074, China

Sustainability, 2025, vol. 17, issue 22, 1-30

Abstract: Railway hazardous chemical transportation, a high-risk activity that endangers personnel, infrastructure, and ecosystems, directly undermines the sustainability of the transportation system and regional development. Traditional risk management algorithms, which rely on empirical rules, result in sluggish emergency responses (with an average response time of 4.8 h), further exacerbating the environmental and economic losses caused by accidents. The standalone DBSCAN algorithm only supports static spatial clustering (with unoptimized hyperparameters); it lacks probabilistic reasoning capabilities for dynamic scenarios and thus fails to support sustainable resource allocation. To address this gap, this study develops a DBSCAN–Bayesian network fusion model that identifies risk hotspots via static spatial clustering—with ε optimized by the K-distance method and MinPts determined through cross-validation—for targeted prevention; meanwhile, the Bayesian network quantifies the dynamic relationships among “hazardous chemical properties-accident scenarios-material requirements” and integrates real-time transportation and environmental data to form a “risk positioning-demand prediction-intelligent allocation” closed loop. Experimental results show that the fusion algorithm outperforms comparative methods in sustainability-linked dimensions: ① Emergency response time is shortened to 2.3 h (a 52.1% improvement), with a 92% compliance rate in high-risk areas (e.g., water sources), thereby reducing ecological damage. ② The material satisfaction rate reaches 92.3% (a 17.6% improvement), and the neutralizer matching accuracy for corrosive leaks is increased by 26 percentage points, which cuts down resource waste and lowers carbon footprints. ③ The coverage rate of high-risk areas reaches 95.6% (a 16.4% improvement over the standalone DBSCAN algorithm), with a 27.5% reduction in dispatch costs and a drop in resource waste from 38% to 11%. This model achieves a leap from static to dynamic decision-making, providing a data-driven paradigm for the sustainable emergency management of railway hazardous chemicals. Its “spatial clustering + probabilistic reasoning” path holds universal value for risk control in complex systems, further boosting the sustainability of infrastructure.

Keywords: railway hazardous chemicals transportation; DBSCAN algorithm; Bayesian network; emergency material dispatch; algorithm fusion (search for similar items in EconPapers)
JEL-codes: O13 Q Q0 Q2 Q3 Q5 Q56 (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
https://www.mdpi.com/2071-1050/17/22/10085/pdf (application/pdf)
https://www.mdpi.com/2071-1050/17/22/10085/ (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:gam:jsusta:v:17:y:2025:i:22:p:10085-:d:1792338

Access Statistics for this article

Sustainability is currently edited by Ms. Alexandra Wu

More articles in Sustainability from MDPI
Bibliographic data for series maintained by MDPI Indexing Manager ().