Algorithm 1: Initialize the Subpopulations Based on Clustering

Input: -Search space R -The number of subpopulations K -Initial population Pop -$\mathrm{Initial} \mathrm{population} \mathrm{individual} {\mathbf{P}\mathbf{o}\mathbf{p}}_i$ -$\mathrm{Cluster} \mathrm{centers} {\mathbf{P}}_{a1},{\mathbf{P}}_{a2},..,{\mathbf{P}}_{aK}$ Output: -Subpopulations subPop1, subPop2, …, subPopK 1: Generate the initial population Pop within the search space R; 2: Randomly select K individuals ${\mathbf{P}}_{a1},{\mathbf{P}}_{a2},..,{\mathbf{P}}_{aK}$ as cluster centers; 3: for each ${\mathbf{P}\mathbf{o}\mathbf{p}}_i$ in Pop: 4:  Calculate the Euclidean distance ${‖{\mathbf{P}\mathbf{o}\mathbf{p}}_i-{\mathbf{P}}_{ai}‖}_2$ between ${\mathbf{P}\mathbf{o}\mathbf{p}}_i$ and each cluster center ${\mathbf{P}}_{ai}$; 5:  Assign ${\mathbf{P}\mathbf{o}\mathbf{p}}_i$ to the subpopulation subPopi corresponding to the nearest cluster center ${\mathbf{P}}_{ai}$; 6: end for 7: Return subPop1, subPop2, …, subPopK