A Genetic Algorithm for the Group-Technology Problem

The design and production planning of cellular manufacturing systems requires the decomposition of a company’s manufacturing assets into cells. The set of machines has to be partitioned into machine-groups and the products have to be partitioned into part-families. Finding the machine-groups and their corresponding part-families leads to the combinatorial problem of simultaneously partitioning those two sets with respect to technological requirements represented by the part-machine incidence matrix. This article presents a new solution approach based on a grouping genetic algorithm enhanced by a heuristic motivated by cluster analysis methods.