A hybrid GA-AUGMECON method to solve a cubic cell formation problem considering different worker skills

• Different worker skills are considered in a three-dimensional cell formation problem.
• Part quality indexes are influenced by different worker skill assumptions.
• Sometimes a more skillful worker is called from an outsider cell to work.
• We configure cells by minimizing movements and maximizing part qualities simultaneously.
• A new hybrid algorithm based on augmented ε-constraint method and genetic algorithm is developed.

Part quality and consequently customer satisfaction besides cost functions are two of the most important issues for any firm. Balancing between these two goals leads to full utilization from manufacturing resources. Formerly, in cubic cell formation problem, where a part on a machine can be processed by various workers, worker assignment was done just by minimizing inter-cell movement criterion; so, the workers assigned into the processing cell are mostly selected rather than outsider workers. But, it is rational for the ties to be broken by skills of different workers in performing a special part on the dedicated machine. In this paper, a bi-objective cubic cell formation is presented with two non-homogeneous objective functions in order to minimize the inter-cell movements and maximize a part quality index. Quality index for each part is represented through a cubic matrix containing integer values of 1–5 (representing very bad, bad, medium, well and very well), which qualifies the process of part on a specific machine by a specific worker. To solve the problem, a hybrid GA-augmented ε-constraint method (GA-AUGMEON) is developed to reduce time consuming difficulty of AUGMECON method. To validate the model and the GA-AUGMECON algorithm, some randomly generated examples in small and large size are solved.