Semiconductor fab system level time-constraint control with uncertainty informed machine learning

Semiconductor fab system level time-constraint control with uncertainty informed machine learning

EVENT DATE
3 Oct 2025
Please refer to specific dates for varied timings
TIME
2:00 pm 3:00 pm
LOCATION
SUTD Think Tank 21 (Building 2, Level 3, Room 2.310)

Semiconductor manufacturing systems, fabs, are the most complex manufacturing systems. They provide an ideal environment for the conception of intelligent production control algorithms. A major complexity driver are time-constraints, that limit the maximum time between two processes. In the state-of-the-art coordination is impossible on system level. A new real-time data-based approach for intelligent production control specifically designed for time-constrained complex job shops is proposed. It uses real-time system replicas, along with uni-/multi-variate ML models and digital twins, to predict time constraint violations. The approach then derives the production control from these violation likelihood predictions that are primarily obtained from ML and their uncertainty quantification. Validated in a real semiconductor fab, it showed significant improvements over traditional methods by preventing violations.

Related papers/articles:

  • May, M. C., Oberst, J., & Lanza, G. (2024). Managing product-inherent constraints with artificial intelligence: production control for time constraints in semiconductor manufacturing. Journal of Intelligent Manufacturing, 1-18.
  • May, M. C., Overbeck, L., Wurster, M., Kuhnle, A., & Lanza, G. (2021). Foresighted digital twin for situational agent selection in production control. Procedia CIRP, 99, 27-32.
  • May, M. C., Kiefer, L., Kuhnle, A., & Lanza, G. (2022). Ontology-based production simulation with ontologysim. Applied Sciences, 12(3), 1608.
  • May, M. C., Albers, A., Fischer, M. D., Mayerhofer, F., Schäfer, L., & Lanza, G. (2021). Queue length forecasting in complex manufacturing job shops. Forecasting, 3(2), 322-338.
  • May, M. C., Behnen, L., Holzer, A., Kuhnle, A., & Lanza, G. (2021). Multi-variate time-series for time constraint adherence prediction in complex job shops. Procedia CIRP, 103, 55-60.
Speaker’s profile

Marvin Carl May received his Bachelor and Masters in Industrial Engineering and Management as well as Information Systems from Karlsruhe Institute of Technology (KIT), Germany. In 2023 he completed his PhD at KIT in Mechanical Engineering focusing on AI in manufacturing. During that time he was a visiting researcher at the National University of Singapore. In 2024 he was awarded the Dr.-Ing. Willy Höfler prize for the best PhD thesis in mechanical engineering, and in 2025 the respective KIT doctoral award, Innovation prize and the Helmholtz association doctoral award for Information (Germany-wide). Subsequently, he stayed at MIT as a PostDoc. Since 2025 he is assistant professor for Industrial AI in Aerospace and Mechanical Engineering at Nanyang Technological University, Singapore.

For more information about the ESD Seminar, please email esd_invite@sutd.edu.sg
 

ADD TO CALENDAR
Google Calendar
Apple Calendar