SIMULTANEOUS IDENTIFICATION OF GROUP TECHNOLOGY MACHINE CELLS AND PART FAMILIES USING A MULTIPLE SOLUTION FRAMEWORK

The problem of grouping parts into families and machines into cells (i .e. the group technology configuration problem or GTCP) has been addre ssed in the literature using a wide variety of statistical, numerical, and optimization techniques. This paper presents an iterative problem modification oriented framework for addressing the GTCP. The proposed framework involves a change in the basic philosophy of developing and testing procedures for solving the GTCP. That is, instead of viewing the process as one where a proposed procedure is applied to a 'whole' problem, from which a solution is produced and evaluated, we conjectur e that the process should involve generating a number of solutions via a broad variety of techniques, and examining them for consistency. To illustrate this framework we present operational and computational re sults on a published problem. Specifically we illustrate its use with a number of heuristic algorithms, focusing on partitioning a bipartite model formulation and the concurrent identification of machine cells and part families.