What is it about?
Many real-world problems involve balancing several conflicting goals, and what counts as a “good” solution depends on what a decision maker actually prefers. In this work, we introduce a new way to evaluate optimization results using two simple preference levels: what is considered fully satisfactory (aspiration) and what is considered unacceptable (reservation). Based on these, we propose indicators that measure how well solutions match these preferences, without conflicting with the usual way solutions are compared in optimization.
Featured Image
Photo by Clay Banks on Unsplash
Read the Original
This page is a summary of: D-PHI: Desirability-Based Hypervolume Indicator for Interactive Multiobjective Optimization Using Aspiration and Reservation Levels as Preferences, ACM Transactions on Evolutionary Learning and Optimization, January 2026, ACM (Association for Computing Machinery),
DOI: 10.1145/3794854.
You can read the full text:
Contributors
The following have contributed to this page
