Support vector clustering algorithm for cell formation in cellular manufacturing systems

Abstract

Support Vector Clustering (SVC) algorithm is presented in this study to resolve the cell creation problem in Cellular Manufacturing Systems, even if prior approaches have been used to solve this problem as well. A Gaussian kernel is used in the SVC technique to translate data points from data space to Hilbert space. The method then performs in Hilbert space for the smallest sphere that contains the data images. When the minimal sphere is transferred back to data space, it breaks up into multiple parts, each of which encloses a different point cluster. The scale, at which cluster formation in the data is examined, is determined by the Gaussian kernel's width. Managing overlapping clusters and outliers is made easier by the soft margin constant. The performance of the SVC algorithm is conducted using a collection of two test problems from the literature of Cellular manufacturing systems. The encouraging results are obtained.