An improved fuzzy C-means algorithm for manufacturing cell formation

This paper presents an improved fuzzy C-means algorithm to solve the manufacturing cell formation problems. The proposed algorithm, which integrates the subtractive algorithm (to produce an initial solution), the fuzzy C-means (FCM) algorithm and a solution selecting procedure (to identify the best solution), remedies the major weaknesses of original FCM clustering. We test the performance of the proposed algorithm with 20 data sets from open literature and 60 generated data sets. Our experiments show that the proposed approach performs much better than the original FCM and the solutions are consistent with the best solutions found in references or the control solutions.