Lecture
Joint Lectures on Evolutionary Algorithms (JoLEA)
- Date
- Wednesday 15 February 2023
- Time
- Address
- Online
Incorporating Decision-Maker’s Preferences into the Automatic Configuration of Bi-Objective Optimisation Algorithms
Automatic configuration (AC) methods are increasingly used to tune and design optimisation algorithms for problems with multiple objectives by means of unary quality indicators, which assign a single scalar value to an approximation to the Pareto front. These quality indicators, however, imply preferences beyond Pareto-optimality that may differ from those of the decision maker (DM). Although it is possible to incorporate DM’s preferences into quality indicators, e.g., by means of the weighted hypervolume (HV) indicator, expressing preferences in terms of weight functions is not always intuitive nor an easy task for a DM, in particular, when comparing the stochastic outcomes of several algorithm configurations. A more visual approach to compare such outcomes is the visualisation of differences in their empirical attainment functions (EAFs). We propose using such visualisations as a way of eliciting information about regions of the objective space that are preferred by the DM and present a method to convert the information about EAF differences into a weighted HV that will assign higher quality values to approximation fronts that result in EAF differences preferred by the DM. The resulting weighted HV may be used by an AC method to guide the configuration of multi-objective optimisers according to the preferences of the DM. We apply our approach to re-configuring, according to different DM’s preferences, a multi-objective optimiser tackling a real-world production planning problem arising in the manufacturing industry.
About Manuel López-Ibáñez
Manuel López-Ibáñez is a Senior Lecturer (Associate Professor) in the Decision and Cognitive Sciences Research Centre at the Alliance Manchester Business School, University of Manchester, UK. His main expertise is on the application of metaheuristics, including local search, evolutionary algorithms and ant colony optimization, to optimization problems, including continuous, combinatorial, and multi-objective problems. His current research is on the experimental analysis and automatic configuration and tuning of stochastic optimization algorithms, in particular, when applied to multi-objective optimization problems. He has also developed several widely used software tools such as irace, hypervolume computation and EAF computation and visualization.
