Potensi Pelanggan Tunggakan PDAM Menggunakan Metode K-Medoids dengan Optimasi Ant Colony Optimization (ACO)

PDAM in carrying out operational activities is greatly influenced by the receivables or arrears of customer water bills. Some factors that influence customer patterns in delinquent water bills are customer class and consumption of water usage, which affects the water bill paid by the customer. This study will apply the K-Medoids clustering method to find out customers who are delinquent in the PDAM by optimizing the selection of cluster centers using the Ant Colony Optimization (ACO) algorithm. In this study the combination of ACO and K-Medoids methods is called ACOMedoids. The results with the ACOMedoids method can produce a high level of accuracy from the comparison of clustering data with actual bill data. This can be seen from the results of accuracy which is always better than the K-Medoids method, which is the highest achieves 97.65% accuracy for ACOMedoids while K-Medoids is 88.29%. Accuracy results show that the ACO algorithm can produce optimal cluster center points in the clustering process of the K-Medoids method.