A CONSTRUCTIVE ALGORITHM FOR CELL FORMATION IN CELLULAR MANUFACTURING SYSTEMS CONSIDERING CELL LOAD VARIATIONS

摘要

Group Technology is a proven strategy in reducing manufacturing lead time and enhancing productivity. Most of the cell formation methods are focused on lowering inter-cell movements. However, consideration of production volume can also reduce the bottlenecks of cells. In this study, an attempt has been made to form part families by developing a genetic algorithm which minimizes the difference between the production volumes of the part families and maximizes the similarities among the parts in the part families. Thus, it minimizes cell load variations along with reducing inter-cell movements. Also, in case of the unavailability of production volume information, the algorithm has the capability of giving results with a part-machine incidence matrix only. This algorithm is combined with a search heuristic to form machine cells. A new concept named 'Identical Machine Replacement' has also been introduced here, and it has shown promising results on reducing operations that fall outside the cluster. The effectiveness of the proposed algorithm is measured with three extremely popular performance metrics: 'Efficiency' by Chandrasekharan and Rajagopalan, 'Efficacy' by Kumar and Chandrasekharan and 'Performance Measure, β' by Ohta and Nakamura. The result is compared with some published methods, and it was possible to obtain up to 11.43% more Efficiency, 19.07% more Efficacy, and 6.12% better Performance. On average, this algorithm resulted in a 2.45%, 4.01%, and 2.67% increase in Efficiency, Efficacy, and β, respectively.