Title: Efficient Search for the Best Trade-off with Bayesian Optimisation
Abstract: Bayesian optimisation is a sample efficient technique for expensive black-box optimisation problems. After a brief introduction to Bayesian optimisation, this talk will explain how it can be applied to problems with multiple objectives. We will look at approaches that try to identify a good approximation of the entire Pareto front, but also approaches that aim to return only a small set of alternatives, such as the knee. Finally, we show how preferences can be learned interactively from pairwise preference elicitations so that the most preferred trade-off solutions can be found more efficiently.
About the speaker: Prof. Jurgen Branke is the Professor of Operational Research and Systems at Warwick Business School. He is visiting the Multiobjective Optimization Group at JYU as a JYU Visiting Fellow. His research interests include nature inspired optimisation, logistics and scheduling, multiobjective optimisation and decision making, optimisation in presence of uncertainty, simulation-based optimisation, and Bayesian optimisation.
