An Adaptive Multi-Population Approach for Sphericity Error Evaluation in the Manufacture of Hemispherical Shell Resonators

. 2024 Feb 28;24(5):1545. doi: 10.3390/s24051545

Algorithm 2: Adaptive Reconstruction and Elimination Mechanism

Input:

-
Search space R
-
Population Pop
-
Subpopulations subPop1, subPop2, …, subPopK
-
Optimal solutions of subpopulations pbest1, pbest2, …, pbestK

Output:

-
The global optimal solution gbest

1: While the number of subpopulations K falls below the threshold:
2: Perform the search process (as described in Section 2.2.2);
3: Utilize the K-means clustering method to reorganize all subpopulations;
4: Select pbest1, pbest2, …, pbestK as the cluster centers from each subPop1, subPop2, …, subPopK;
5:

for each individual {P o p}_{i}

in Pop:
6:

Calculate Euclidean distance {‖{P o p}_{i} - p b e s t K‖}_{2}

between {P o p}_{i}

and each pbest1, pbest2, …, pbestK;
7:

Assign {P o p}_{i}

to the subpopulation with the closest distance to its corresponding cluster center;
8: end for
9: for each subPopK:
10: if the number of individuals in a subpopulation is less than the Threshold:
11: Eliminate the subpopulation subPopK;
12: end if
13: end for
14: Continue the search process;
15: end while
16: Return the global optimal solution gbest.