Case study · October 24, 2023

Benchmark of PI-BO by ZF Friedrichshafen at the NAFEMS seminar on AI and machine learning in CAE-based simulation

ZF Friedrichshafen presented benchmark results for PI-BO, the Bayesian optimization from the STOCHOS software package, at the NAFEMS seminar "Artificial Intelligence und Machine Learning in der CAE-basierten Simulation" in October 2023.

Benchmark table from the ZF presentation comparing PI-BO, Adaptive MOP, and DIM-GP LHS by R squared and MAPE at 100, 500, and 1000 samples, for three model variants with 22, 46, and 94 free parameters, PI-BO rows highlighted.
ZF comparison of adaptive sampling versus LHS across sample set sizes (table in German).

The seminar was a very interesting and informative event, with a variety of companies and scientists showing state-of-the-art high-fidelity methods for highly complex CAE applications.

PI Probaligence contributes to the future vision of advanced machine learning trained with small data to predict simulation outcomes with high efficiency and accuracy. See the recently published results from ZF, using our PI-BO from the STOCHOS software package.

Next step
Method
Bayesian Optimization

How the search picks each next run from the last result.

Guide
Bayesian optimization for engineers

Reach a better design in tens of runs, not hundreds.

Interactive demo
Try to beat STOCHOS

Three use cases, about two minutes each.

NewerWe at PI believe that highly effective, specialized A... OlderGreen Tech kick-off event for project Na, Logisch

← All news and guides

Partners, customers, and research collaborators
AnsysCADFEMSimuTech GroupMEScoTSNEBoschZFGEMUDLRAdler LackeMankiewiczDuluxPlixxentFraunhoferHochschule NiederrheinFUELL Lab AutomationHumotionUniversitaet HamburgRobert Bosch Stiftung