[4M1-GS-2a] Machine learning

10:15 AM - 10:30 AM JST(1:15 AM - 1:30 AM UTC)

[4M1-GS-2a-06] Revisiting the Definition of a Region of Interest in Preference-based Evolutionary Multi-objective Optimization

〇Ryuichi Mogami¹, Ryoji Tanabe¹ (1. Yokohama National University)

Keywords:

Evolutionary Computation

Multi-objective optimization aims to obtain a solution set that simultaneously minimizes multiple objective functions. Evolutionary multi-objective optimization (EMO) is an effective approach for finding an approximation of the whole Pareto front. Unlike EMO, preference-based EMO (PBEMO) aims to approximate a region of interest (ROI) defined by the preference information from a decision maker. However, the ROI has been loosely defined in the EMO community, especially when the objective functions have different scales. To address this issue, this paper analyzes the influence of normalization of objectives on the ROI. We demonstrate that the ROIs defined in the normalized and original objective spaces can differ from each other. Our results also show the difficulty in approximation the ROI defined in the normalized objective space.