Working Group 4

Optimisation Under Uncertainty

Leader: Per Kristian Lehre, UK
Vice-Leader: Michael Hellwig, AT

WG4 brings together researchers from statistical modelling and optimisation to advance problem definitions and methods related to uncertainty handling. It aims to improve and develop optimisation algorithms for classes of uncertainties commonly arising in real-world applications.

Tasks

• Structuring and organising the knowledge domain with respect to the type of uncertainty (aleatoric, epistemic and deep uncertainty) and the different ways to model uncertainty (including interval arithmetic, probabilistic, and fuzzy-logic approaches, and interesting new directions such as evidence-based reasoning). This requires collaboration with WG1 and WG2. Orthogonal to this classification, uncertainty occurs in different parts of optimisation problem solving, such as in the statistical modelling of objective functions, in noisy objective functions, in constraint-handling or in the preferences provided by the decision maker (collaboration with WG3).
• Developing and evaluating improved algorithms for optimisation under uncertainty of continuous and discrete/combinatorial optimisation problems (collaboration with WG2) involving one or more objectives (collaboration with WG3), while considering the specific challenges of benchmarking algorithms for problems involving uncertainty (collaboration with WG6).

Latest news