Component grouping for GT applications - a fuzzy clustering approach with validity measure

The variety of the currently available component grouping methodologies and algorithms provide a good theoretical basis for implementing GT principles in cellular manufacturing environments. However, the practical application of the grouping approaches can be further enhanced through extensions to the widely used grouping algorithms and the development of criteria for partitioning components into an 'optimum' number of groups. Extensions to the fuzzy clustering algorithm and a definition of a new validity measure are proposed in this paper. These are aimed at improving the practical applicability of the fuzzy clustering approach for family formation in cellular manufacturing environments. Component partitioning is based upon assessing the compactness of components within a group and overlapping between the component groups. The developed grouping methodology is experimentally demonstrated using an industrial case study and several well known component grouping examples from the published literature.