Skip to content
Open access

Region-Focused Evolutionary Search for Dynamic Multi-Objective Optimization

Oct 2026 · IEEE/CAA Journal of Automatica Sinica · Vol 13, pp. 2378-2392 · 0 citations · 58 references

Abstract

The growing requirements of solving dynamic multi-objective optimization problems (DMOPs) have increased the need to design dynamic multi-objective evolutionary algorithms (DMOEAs). As a powerful tool for solving DMOPs, DMOEAs based on prediction mechanisms have attracted more and more attention. However, limited efforts have been made to efficiently and accurately predict the changing Pareto optimal set (POS) in the local space. Taking this cue, we devise a novel region-focused evolutionary directional prediction method, called the knowledge-transfer grid-based DMOEA (KG-DMOEA), to accurately predict evolutionary directions and accelerate population iteration speed. First, a hierarchical detection technique is designed, which can select individuals on a global scale to improve detection accuracy. Second, a grid-based regionalized evolutionary direction prediction method is proposed, which divides individuals into different grids and improves change response accuracy through region-focused evolution strategies. Third, a knowledge transfer strategy is embedded into the algorithmic framework, enabling enhanced population convergence and driving individuals to rapidly evolve toward the optimal Pareto front. Furthermore, different test problems are adopted to verify the devised strategy where the comparative results demonstrate that the devised novel approach can efficiently reduce the time cost while improving the quality of solutions over existing algorithms.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.