An efficient dominance decomposition-based deep graph evolutionary algorithm for the expensive multi-objective optimization

doi:10.1016/j.eswa.2026.131379

An efficient dominance decomposition-based deep graph evolutionary algorithm for the expensive multi-objective optimization

2026-01-29

0

PRE

AI

DOI：10.1016/j.eswa.2026.131379

Original

Original request for help

Save

Abstract

En 中文

• Proposes a dominance decomposition-based deep graph evolutionary algorithm. • Decomposes complex dominance prediction into sub-objective superiority learning. • Introduces graph neural networks into expensive multi-objective optimization. • Improves efficiency via multi-task surrogate modeling and cluster filtering. • Demonstrates strong performance on benchmark and real-world optimization tasks.

Keywords:

dominance decomposition

graph neural networks

multi-objective optimization

surrogate modeling

evolutionary algorithm

An efficient dominance decomposition-based deep graph evolutionary algorithm for the expensive multi-objective optimization

Abstract

Journal

Organization