An ART1 based algorithm for Cell Formation with sequence data

Modern batch type production industries prefer cellular layout for achieving reduced lead time, reduced setup time, reduced inventory, reduced material handling time and increased machine utilization effectively. The products are grouped into part families based on similarities in manufacturing and design attributes, and the machines are allocated into machine cells to produce the identified part families so that productivity and flexibility of the system can be increased. Usually the zero-one part-machine incidence matrix is the common input to any clustering algorithm which represents the part visiting a particular machine as one and zero otherwise. The major demerit with such models is that the real life production factors are not accounted for. In this paper, the production sequence of the parts is also considered to enhance the quality of the solution obtained for the cell formation (CF) problem by the proposed algorithm, which is based on adaptive resonance theory (ART) networks. The performance of the proposed algorithm is tested with example problems and the results are compared with the existing methods found in the literature. The performance measures considered are the group technology efficiency and the number of inter-cell movements. The results presented clearly shows that the performance of the proposed ART1 based algorithm is comparable over the other methods for smaller size problems and better for large size problems.