Dr. Holger Brandl
Analytics Solution Architect
Optimizing semiconductor production using AI and scheduling to balance competing goals
The presented work is part of the international research project AISSI (https://aissi-project.com/), where SYSTEMA has teamed up with industrial manufacturing partners (Bosch & Nexperia), academic patterns (KIT), and modelling experts (D-SIMLAB), to develop, integrate and apply novel AI-based approaches to semiconductor execution planning. By embedding reinforcement learning in a continuous framework for autonomous, integrated production and maintenance scheduling, we strive to outperform current state-of-the-art approaches in terms of efficiency and cost-effectiveness.
This interdisciplinary team is constantly seeking ways to improve the efficiency, automation and optimization of semiconductor production. Using advanced methods such as meta-heuristics, simulation, constraint solving, and digital twin technology, we have created a testbed to evaluate different execution planning methods. By utilizing a digital twin to monitor and analyze data, we can efficiently reveal additional production potential with high accuracy.
Together with our application partner NEXPERIA, we are comparing various optimization techniques, including meta-heuristics, combinatorial optimization, and reinforcement learning to find the best balance between conflicting objectives in a high-throughput wafer fabrication area. Our goal is to provide line engineers with a clear and reliable method for scheduling and verifying production processes.