Session 4 – 22 June

Towards full factory scheduling

15:30 – 16:00

The ordering in which heterogeneous products are manufactured on parallel machines can have a substantial impact on the efficiency of a factory. With only 60 waiting jobs, there are already more possibilities to arrange them than there are atoms in the universe – way too many for humans to explore. Therefore, Bosch’s semiconductor factory in Reutlingen used to be controlled with dispatching rules written by human experts. While rules have certain advantages, they’re known to be inefficient in comparison to scheduling approaches when it comes to optimally utilizing the available production capacity. With scheduling, it’s also possible to consider constraints that were previously ignored due to their complexity and to formulate goals explicitly. This talk covers the successes and challenges of moving a semiconductor factory entirely controlled by rules towards a factory using scheduling based on state-of-the-art combinatorial optimization techniques.

Sebastian Bayer is a data scientist and AI consultant at the Robert Bosch Center for Artificial Intelligence where he works on solving challenging problems of Bosch-internal customers. He’s currently concentrating on topics involving mathematical optimization and statistical techniques and modeling. For the last three years, he’s been working in collaboration with the Robert Bosch Wafer and Sensor Fab in Reutlingen, where he’s replacing rule-based dispatching with optimized scheduling routines. He holds an MSc in economics and a PhD in econometrics from the University of Konstanz, where he focused on applying statistical and mathematical models on modeling financial market risks.

Sebastian Bayer

Data scientist at Bosch Center for Artificial Intelligence