A data mining based clustering approach to group technology

Cellular manufacturing is an essential application of group technology (GT) in which families of parts are produced in manufacturing cells. This paper describes the development of a cell formation approach based on association rule mining and 0-1 integer programming. It is valuable to find the important associations among machines such that the occurrence of some machines in a machine cell will cause the occurrence of other machines in the same cell. A clustering model using the discovered association data is formulated to maximize the closeness measures among machines within each cell. From the results of three medium-sized problems, the proposed approach shows its ability to find quality solutions of cell formation problems.