.png){kind=icon}
.png){kind=icon}
Return
Stochastic production scheduling in modular construction: A digital twin case study of a wood framing machine
.png){kind=icon}
PRE
AI
D
J
A
Y
M
S
DOI：10.1016/j.aei.2026.104737.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Off-site wood construction continues to face major challenges in production planning due to unpredictable human–machine interactions and non-linear workflow variability. This study introduces the Wood-framing Simulation-based Prediction Digital Twin (WSP-DT), a framework that integrates discrete-event simulation, real-time Supervisory Control and Data Acquisition (SCADA) monitoring, and predictive analytics to improve scheduling reliability. WSP-DT advances existing digital twin frameworks by integrating empirically validated stochastic models of human–machine interaction with rule-aware scheduling constraints, which is demonstrated on a physical wood framing machine station with 47 wall panels. A three-scenario case study was conducted to evaluate compliance, workload balance, and resource utilization. Scenario 1 represents a speed-optimized baseline with overtime tolerance, Scenario 2 enforces strict regulatory compliance with standard work hours, and Scenario 3 implements 24-hour operations balancing speed and compliance. An operational efficiency score was used to measure balanced daily capacity utilization, penalizing overload periods and rewarding stable pacing. Results show that scenario 2 eliminates all regulatory violations and achieves an operational efficiency score of 62.7%, compared to 13.3% in scenarios 1 and 3. Although scenario 2 substantially increases project duration, it eliminates all overload and night-work conditions, highlighting the non-linear trade-off between compliance and production speed. The digital twin achieved a duration error of 0.86%, confirming high predictive accuracy. An ablation analysis confirmed that removing individual WSP-DT components degrades prediction error by 6.6-9.2 ×, compliance efficiency by 4.7 ×, and temporal fidelity by 5 ×, validating the integration architecture as the primary source of system performance. The findings highlight that sustainable scheduling in modular construction depends not on maximum speed but on balanced optimization among speed, compliance, and resource stability.
Keywords:
Digital Twin
Stochastic Production Scheduling
Modular Construction
Human–Machine Interaction
Predictive Analytics
Journal
IF:
9.9
Papers:
4.0K
Citations:
1.7W
Organization
Cited Papers