Machine-cell and Part-family Formation in Cellular Manufacturing Using a Two-phase Clustering Algorithm ⋆

This paper presents a two-phase clustering algorithm for machine-cell and partfamily formation in the design of cellular manufacturing systems. The proposed algorithm begins with the determination of initial cluster centers via a linear assignment method using the least similar group representatives in its first phase. A fuzzy C-means clustering method is followed in its second phase for part-family and machine-cell formation using the obtained initial cluster centers. The two-phase algorithm can remedy the problem of clustering inconsistency resulting from the fuzzy C-means method with random initializations. Experimental results on many benchmark data sets based on multiple performance criteria substantiate the effectiveness of the proposed algorithm.