A Novel Two-Phase Cooperative Co-evolution Framework for Large-Scale Global Optimization with Complex Overlapping

Cooperative Co-evolution, by decomposing the problem space, is a key approach for large-scale global optimization. While it outperforms non-decomposition algorithms when subspaces are disjoint, overlapping variables complicate decomposition and degrade performance. To address this, we propose a novel two-phase cooperative co-evolution framework for complex overlapping problems, incorporating an effective method for decomposing overlapping problems based on their mathematical properties. We also introduce a customizable benchmark to extend existing ones for experimentation. Extensive experiments show that our framework significantly outperforms existing algorithms, revealing the characteristics of overlapping problems and the strengths of cooperative co-evolution and non-decomposition algorithms. The work is open-source at: https://github.com/GMC-DRL/HCC.