A directional decomposition heuristic for one-dimensional, non-equidistant machine-cell location problems

Abstract

Machine-cell location (MCL) problems have been major points of interests of many researchers dealing with cell formation and material flow analysis in a manufacturing shop floor. Since the internal configuration of the production line in a cell is the outcome of both machine-group formation and machine-location problems, the locations of these cells along a material-handling track (or transporter path) are of further interests for the refinement of the results. As the processing technology improves and computational capabilities enhance, the scope of this assignment is even more captivated by potential total benefit. In this research, a one-dimensional, non-equidistant MCL problem is considered, where the input of the system starts from the output of a group formation problem. First the bi-directional movements are decomposed into backward and forward gains and these incremental gains are exploited to explore the most potential search procedure in either of the decomposed directions. The solution procedure is simple, but efficient and good with respect to both time and quality of solution. This directional decomposition heuristic generates a relatively good solution as compared to other existing solutions in MCL analysis.