Algorithms for machines and parts grouping to design cells

Abstract

This thesis is concerned with the cell formation in cellular manufacturing system which has various advantages in manufacturing facilities with small batch sizes and high variety of products. Different cell formation methods are proposed considering the fact that they can be varied depending upon the input data. First, we formulate two different mathematical models, the p-median model and the quadratic assignment model, for the problem of machine cell formation based on the incidence relationships between machines and parts. A heuristic method for the quadratic assignment model is also presented. Secondly, we deal with the cell formation problem with alternative part processing routes. Two inherent problems, the route selection problem and the part families formation problem are solved separately in the proposed two stage procedure. At the first stage, the compatibility coefficient is introduced for the purpose of route selection. Finally, we deal with the cell formation problem with alternative part processing routes considering the production and cost related restriction explicitly. Again, a two stage procedure in which the route selection is handled at the first stage with the objective of minimizing the production and machine related cost and the part families formation is dealt with at the second stage with the objective maximizing similarities in the same cells.