Articles on Crash and Vehicle Control Dynamics Accepted at the Genetic and Evolutionary Computation Conference (GECCO)
GECCO is a leading conference in the field of optimization algorithms gleaned from principles of natural evolution. The conference takes place in Boston, MA, from July 9 –13. Many world-leading experts in the field, PhD students and industry experts will come together to discuss and advance the state-of-the-art in the field. We are very happy that our joint work with BMW group within the newAIDE project has been accepted for presentation and publication at the conference.
The paper “Learning the Characteristics of Engineering Optimization Problems with Applications in Automotive Crash” by Fu Xing Long (BMW), Bas van Stein (Leiden University), Moritz Frenzel (BMW), Peter Krause (divis intelligent solutions GmbH), Markus Gitterle (Munich University of Applied Sciences) and Thomas Bäck (Leiden University) focuses on learning characteristics of engineering optimization problems that make them different from other optimization tasks. The result shows that there are strong differences, such that it makes sense to develop specific optimization methods for engineering design optimization problems. We are proud that this work has been nominated for a best paper award.
The paper “One-Shot Optimization for Vehicle Dynamics Control Systems: Towards Benchmarking and Exploratory Landscape Analysis” by André Thomaser (BMW), Anna Kononova (Leiden University), Marc-Eric Vogt (BMW) and Thomas Bäck (Leiden University) looks at control systems for vehicle dynamics. For exploiting parallelism, one-shot approaches for optimization are investigated and combined with surrogate models to predict a high-quality solution, which is then validated using the simulator. The results show that such an approach works very well with random forest and support vector machine surrogates.