Efficient preference learning algorithm for interactive evolutionary multi-objective optimization

doi:10.1016/j.swevo.2025.102254

Efficient preference learning algorithm for interactive evolutionary multi-objective optimization

2025-12-10

0

PRE

AI

DOI：10.1016/j.swevo.2025.102254

Original

Original request for help

Save

Abstract

En 中文

• We introduce a novel evolution framework for diverse preference-based models. • The framework uniquely balances elasticity, efficiency, and distributional quality. • The method outperforms baseline approaches in robustness, efficiency, and success rates. • The framework leads to better-distributed models and superior objective space coverage. • Ensuring uniform model distribution lets avoiding biased solution assessment.

Keywords:

Preference learning

Pairwise comparisons

Interactive procedures

Evolutionary multi-objective optimization

Uniform sampling

Efficient preference learning algorithm for interactive evolutionary multi-objective optimization

Abstract

Journal

Organization