K-Means Centroid Optimization with Genetic Algorithm for Clustering Micro, Small, Medium Enterprises in Yogyakarta

K-Means is a widely used data clustering algorithm due to its simplicity and fast performance. However, the weakness of K-Means is in determining the cluster centroid randomly, which can result in suboptimal clustering results, especially since it tends to get stuck on local solutions. This research aims to overcome this weakness by integrating the Genetic Algorithms (GA) into the K-Means process, optimizing the initial centroid, and improving clustering quality. The method combines GA with K-Means on MSME data in Yogyakarta, where GA rearranges the cluster's initial centroid more optimally. The results showed that this method reduced the average value of the Davies-Bouldin Index (DBI) from 1,819 in conventional K-Means to 1,349 with GA integration, indicating an improvement in cluster quality by 25.9%. These results prove that integration of GA with K-Means improves clustering accuracy and improves cluster separation, as measured by a significant decrease in DBI value