Pascal lecture 2017: Multi-objective optimization meets data science
- Prof. dr. Heike Trautmann, University of Münster
- 27 March 2017
- Snellius Building
Niels Bohrweg 1
2333 CA Leiden
Lecture and drinks
You are cordially invited to attend this lecture and the drink afterwards, which will take place in the Foobar at the ground floor of the Snellius Building.
This talk gives an overview of Trautmann's current research topics and projects in order to foster collaborations with LIACS during the upcoming year.
Research on multiobjective (evolutionary) optimization deals with the integration of experts' knowledge and preferences, understanding problem characteristics such as multimodality of the landscape and designing specific algorithms for multiobjective vehicle routing problems.
Automated algorithm selection, i.e. predicting the best suited optimization algorithm for an optimization problem at hand, is investigated for the Travelling Salesperson Problem as well as continuous black-box optimization with promising perspectives for generalization to the multiobjective context.
Specific focus is on current data science projects such as customer relationship management, monitoring customer service via Facebook and Twitter, related research on stream clustering algorithms and propaganda detection in online and social media.
Prof. Trautmann is hosted by prof. dr. Thomas Bäck and dr. Michael Emmerich.