Improvement in Computation Time of the Finite Multipole Method by Using K-Means Clustering

K-means algorithm is one of the most widely used clustering algorithms in unsupervised machine learning. In the K-means clustering, input data are classified into the corresponding cluster based on the nearest mean criteria. We utilize the K-means clustering to create groups of basis functions in the finite multipole method. By replacing the conventional cube grouping with the K-means clustering, convergence rate in an iterative solver is improved. The computation time is thus alleviated without loss of the calculation accuracy. The performance of the proposed scheme is demonstrated by using both canonical and realistic computer-aided design models.