PERFORMANCE OF SELECTED PART-MACHINE GROUPING TECHNIQUES FOR DATA SETS OF WIDE-RANGING SIZES AND IMPERFECTION

This study addresses the part-machine grouping problem in group techno logy, and evaluates the performance of several cell formation methods for a wide range of data set sixes. Algorithms belonging to four class es are evaluated: (1) array-based methods: bond energy algorithm (BEA) , direct clustering analysis (DCA) and improved rank order clustering algorithm (ROC2); (2) nonhierarchical clustering method: ZODIAC; (3) a ugmented machine matrix methods: augmented p-median method (APM) and a ugmented linear clustering algorithm (ALC); and (4) neural network alg orithms: ART1 and variants: ART1/KS, ART1/KSC, and Fuzzy ART. The expe rimental design is based on a mixture-model approach, utilizing replic ated clustering. The performance measures include Rand Index and bond energy recovery ratio, as well as computational requirements for vario us algorithms. Experimental factors include problem size, degree of da ta imperfection, and algorithm tested. The results show that, among th e algorithms applicable for large, industry-size data sets, ALC and ne ural networks are superior to ZODIAC, which in turn is generally super ior to array-based methods of ROC2 and DCA.