A multidimensional scaling algorithm for group layout in cellular manufacturing

Component families and machine cells formed in a cellular manufacturing system generally contain exceptional operations which introduce intercell movement. A layout problem is defined so as to minimize the material handling consequent to intercell moves. An ordinally ranked distance function is computed on the basis of this, and the problem is recast as a classical problem in non-metric multidimensional scaling (MDS). The alternating least squares scaling method used in the analysis of perception data is employed for solving the problem. An illustrative example is solved in which a two-dimensional and a one-dimensional layout are obtained. The problem is also solved using one of the popular heuristic methods for layout design. It is found that the MDS formulation gives better results.