Multi-objective genetic algorithm for the assembly line worker assignment and balancing problem: A case study in the automotive supply industry

Assembly lines have a significant role in the manufacturing sector due to high production quantities and industrialisation. The primary problem in assembly line design is the improper allocation of tasks and workers to their respective workstations. This problem can generally be divided into two classes. The first one is to minimise the number of workstations required to achieve a specific cycle time, while the second is to minimise the cycle time with limited workers. The assembly line worker allocation and balancing (ALWAB) problem is a new type of assembly line balancing problem that has recently gained popularity in academic circles and arises in real assembly lines. Moreover, as the capability and performance of each worker may vary, the workers' task times can also be variable. In this study, the multi-objective ALWAB Type-2 problem, which considers both cycle time and square load allocation objectives, is addressed. A weighted-sum scalarisation-based genetic algorithm (GA) approach is proposed to solve the problem. The thesis was applied in an automotive supplier industry that produces cable assemblies. Since the parameter values significantly affect the performance of the algorithm, a full factorial experimental design is employed for parameter calibration. The proposed GA parameters were compared with the results obtained from irace and the appropriate GA parameters proposed by Mutlu, Polat and Supciller (2013) for the same problem type. The full factorial experimental design method yielded better results. Additionally, a sensitivity analysis was performed by examining the effect of different relative weight values of the objectives on the results. The proposed multi-objective model has been demonstrated to generate reasonable solutions in a reasonable amount of time for a real-life problem.