Model-dependent task assignment in multi-manned mixed-model assembly lines with walking workers

Due to mass customization and extensive market changes, manufacturing companies seek to enhance the flexibility and reconfigurablility of their assembly lines. For instance, to adjust and adapt the line’s capacity to different products and production requirements, workers may move along the stations, or the tasks may be re-assigned. This paper studies the impact of model-dependent task assignment, workforce reconfiguration, and equipment duplication in mixed-model assembly lines. The studied line is paced, and it can process different product models with different sets of tasks and precedence relations. Task and worker assignments to stations may change in each takt, and the goal is to design a line able to handle a predefined set of situations corresponding to different flows of products entering the line. The paper provides a new Mixed Integer Linear Programming (MILP) formulation to minimize the workforce and equipment costs in mixed-model assembly lines with model-dependent task assignment. We provide an efficient reformulation of the MILP by relying on the dualization approach commonly used in robust optimization. In addition, we employ a constructive matheuristic (CM) and a fix-and-optimize heuristic (FOH) to deal with large-scale instances. Extensive computational experiments performed with well-known benchmarks from the literature show that the suggested approaches perform well in terms of solution quality and computational time. In addition, the results reveal that model-dependent task assignment reduces significantly the equipment cost and the number of workers when compared to the classical mixed-model assembly lines with fixed task assignment and walking workers.