Model Predictive Control for Dynamic Resource Allocation in Multi-Class Customs Inspection Systems
Year: 2026
Type: conference-paper
Abstract: Global customs clearance operations face critical challenges in balancing inspection efficiency with security requirements, where stochastic cargo arrivals and limited resources lead to substantial congestion and economic losses. Existing approaches—including static sensor deployment, myopic feedback control, and offline scheduling methods—fail to address the dynamic nature of multi-class inspection demands with heterogeneous service requirements. To bridge this gap, we propose a Model Predictive Control framework for Dynamic Resource Allocation in customs inspection systems (MPC-DRAC). Our approach models the clearance process as a multi-class M/M/c queueing network with risk-stratified service channels, employs rolling horizon optimization to dynamically allocate inspection resources, and incorporates arrival rate forecasting for proactive congestion prevention. Comprehensive simulations under time-varying demand and surge conditions demonstrate that MPCDRAC achieves$\text{3 4. 6 - 7 5. 3 \%}$reductions in average waiting time and$\text{4 8. 9 - 7 7. 3 \%}$reductions in maximum queue length compared to baseline strategies. This framework provides a principled approach to intelligent customs resource management with broader applications to multi-class service systems.
Authors: Rong Hu, Hanchi Sun, Sen Liu
Institutions: Shanghai Customs College, Shanghai Dianji University
Language: English
Field: Decision Sciences
Domain: Social Sciences
Open Access status: hybrid
OpenAlex ID: W7165731442