TRIZ is a tried-and-tested method that engineering teams use to solve tricky technical problems – particularly when two objectives appear to be at odds with one another: for example, if you make a component more stable, it often becomes heavier as well. TRIZ helps to systematically find clever solutions to such conflicting objectives. The catch is that applying the method correctly requires a great deal of experience and specialised training, which most teams lack.
This is precisely where the Master’s thesis comes in. Jonas Wenk has developed an AI system that guides engineers step by step through this process – much like an experienced mentor who asks the right questions at the right time and ensures that no important step is overlooked. Instead of a single AI, five specialised AI modules work together, each responsible for a part of the analysis. It was important to him that the final decision always remains with the human: the system makes suggestions, but only what the human expert confirms is saved.
A thesis with real-world impact
What makes this project special is that, in practice, the system was tested against real-world industrial problems and compared with the work of a certified TRIZ expert – and no methodological errors were found. It also delivered measurably better results than a simpler approach.
The thesis has also already resulted in two academic publications, both of which will be published by Springer Nature: a paper for the TRIZ Users’ Day, co-authored with his supervisors Timo Baumann and Jürgen Moosburger, and a further paper that has been accepted for the TRAI26 conference and has already passed peer review.
The thesis was carried out in collaboration with ams OSRAM, supervised by Prof. Dr Timo Baumann and, externally, by Dr Jürgen Moosburger (ams OSRAM).
