One of the basic elements in cellular manufacturing system design is to setup machines and parts into families. In an ideal case, a complete block diagonal matrix with mutually independent machine-component groups can be identified. However, as the number of parts requiring operations in more than one machine cell (exceptional parts) increases, the effectiveness of the corresponding cellular manufacturing system decreases. This is due to intercellular material handling costs associated with exceptional parts and necessary adjustment in the cellular manufacturing system to accommodate the processing of these exceptional parts.
Many algorithms have been developed for grouping machines and parts into families. Since most of these algorithms are heuristic methods, they do not use any mathematical optimization. For this reason, a number of grouping measures have been developed to evaluate the efficiency of these clustering algorithms. These grouping measures are bond energy, grouping efficiency (GEF), grouping capability index (GCI), grouping efficacy (GC), grouping measure, and the quality index (QI). In this study, the effectiveness of the existing measures in predicting the performance of a cellular manufacturing system is evaluated and the development of a new grouping measure called Fluency Quality (FQ) is presented. The FQ is a measure of the effectiveness of a cell design, and can be used as an objective function for all optimization programs, such as genetic algorithms when designing a cellular manufacturing system.