A configuration-based clustering algorithm for family formation

In Cellular Manufacturing literature, early work focused on the use of routers as a way of forming product families and manufacturing cells. Later, several measures of similarity among machines and parts have been proposed by different authors. In this paper, a new clustering algorithm is proposed that considers the number of machines of each type and the most recent configuration of cells in revising the values of similarity coefficient. The potential benefits of the procedure is demonstrated with a simple example.