| One of the major steps in designing cellular manufacturing systems is to form cells. This involves identification of machine groups and part families. This thesis presents a new approach for optimally grouping the original set of machines into distinct uncorrelated cells. The approach involves two phases. In the first phase, machine cells are formed by applying factor analysis to the matrix of similarity coefficients. In the second phase, integer programming is used to assign parts to the cells. Because of its mathematical basis, the approach assures an optimal cell formation through the generation of uncorrelated cell clusters and the property of variance maximization. Using data set examples, the approach was found to perform very well in terms of a number of objective criteria. Moreover, the approach compares favorably to well-known existing methods. Other advantages such as efficient computational performance, and flexibility during cell formation make this approach very powerful. |
