Pre-variable sorting-based differential grouping for large scale optimization: An interdependency method

This study presents a novel pre-variable sorting-based differential grouping (Pv-DG) method for decomposing large-scale optimization problems into smaller, more manageable sub-problems. By sorting design variables according to their impact on the objective function and applying a recursive differential grouping strategy, the method effectively identifies variable interdependencies. It organizes dependent variables into cohesive groups while distributing independent variables into smaller subgroups. This approach reduces computational complexity and enhances optimization performance. Experimental results of IEEE CEC’2010 and CEC’2013 benchmark functions demonstrate that Pv-DG outperforms four state-of-the-art techniques: differential grouping DG2, fast interdependency identification and recursive differential grouping achieving higher accuracy and efficiency in variable grouping. When integrated within a cooperative co-evolutionary framework, Pv-DG further improves the overall optimization effectiveness for large-scale optimization problems.