Solving Generalized Grouping Problems in Cellular Manufacturing Systems Using a Network Flow Model

This paper focuses on the generalized grouping problem in the context of cellular manufacturing systems (CMS), where parts may have more than one process route. A process route lists the machines corresponding to each part of the operation. Inspired by the extensive and widespread use of network flow algorithms, this research formulates the process route family formation for generalized grouping as a unit capacity minimum cost network flow model. The objective is to minimize dissimilarity (based on the machines required) among the process routes within a family. The proposed model optimally solves the process route family formation problem without pre-specifying the number of part families to be formed. The process route of family formation is the first stage in a hierarchical procedure. For the second stage (machine cell formation), two procedures, a quadratic assignment programming (QAP) formulation, and a heuristic procedure, are proposed. The QAP simultaneously assigns process route families and machines to a pre-specified number of cells in such a way that total machine utilization is maximized. The heuristic procedure for machine cell formation is hierarchical in nature. Computational results for some test problems show that the QAP and the heuristic procedure yield the same results.

💡 Research Summary

The paper tackles the generalized grouping problem in cellular manufacturing systems (CMS), where each part may have multiple alternative process routes. Traditional grouping approaches assume a single route per part, which limits their applicability in modern flexible production environments. To overcome this limitation, the authors propose a two‑stage hierarchical framework.

In the first stage, they formulate the formation of process‑route families as a unit‑capacity minimum‑cost network flow problem. Each possible process route for a part is represented by a pair of nodes, and directed arcs connect these nodes with costs equal to the dissimilarity (i.e., the number of differing machines) between routes. The unit‑capacity constraint forces each route to be assigned to exactly one family, while the flow conservation constraints guarantee that the families are internally coherent. Because the model is a classic minimum‑cost flow, it can be solved to optimality using polynomial‑time algorithms such as the successive shortest‑path or cost‑scaling methods, eliminating the need to pre‑specify the number of families.

The second stage addresses machine‑cell formation and the assignment of the previously created families to a fixed number of cells. Two solution approaches are presented. The first is an exact quadratic assignment programming (QAP) formulation that maximizes total machine utilization by jointly assigning families and machines to cells. Although QAP is NP‑hard, the test instances are small enough that commercial solvers can obtain optimal solutions. The second approach is a hierarchical heuristic: families are first clustered based on similarity, then cells are built iteratively by assigning the most compatible machines. Computational experiments on a suite of benchmark problems show that the heuristic reproduces the exact QAP solutions while requiring far less computational effort.

Results demonstrate that (1) the network‑flow model yields optimal process‑route families without any a‑priori family count, (2) the QAP and heuristic produce identical cell configurations, confirming the heuristic’s effectiveness, and (3) the overall framework is capable of handling the added complexity introduced by multiple routes per part.

The authors acknowledge several limitations. The model assumes deterministic process routes and machine requirements, ignoring stochastic processing times, dynamic routing decisions, and real‑time order variability that are common in actual shop floors. They suggest future extensions to incorporate stochastic parameters, dynamic re‑allocation, and scalability to large‑scale industrial settings. Moreover, they discuss how the network‑flow solution could be embedded within a Cognitive Digital Twin (CDT) architecture, providing real‑time decision support, what‑if analysis, and adaptive reconfiguration in an Industry 5.0 context.

In summary, the paper contributes a novel application of unit‑capacity minimum‑cost network flow modeling to the generalized grouping problem, integrates it with both exact and heuristic cell‑formation methods, and validates the approach through extensive computational testing. This work advances both the theoretical understanding of grouping under multiple routing alternatives and offers a practical, implementable tool for manufacturing planners seeking to design efficient, flexible cellular layouts.