Balancing and scheduling assembly lines with human-robot collaboration tasks

Amir Nourmohammadi, Masood Fathi & Amos H.C.Ng (2022). Balancing and scheduling assembly lines with human-robot collaboration tasks. Journal article in: Computers & Operations Research Volume 140, April 2022.

In light of the Industry 5.0 trend towards human-centric and resilient industries, human-robot collaboration (HRC) assembly lines can be used to enhance productivity and workers’ well-being, provided that the optimal allocation of tasks and available resources can be determined. This study investigates the assembly line balancing problem (ALBP), considering HRC. This problem, abbreviated ALBP-HRC, arises in advanced manufacturing systems, where humans and collaborative robots share the same workplace and can simultaneously perform tasks in parallel or in collaboration. Driven by the need to solve the more complex assembly line-balancing problems found in the automotive industry, this study aims to address the ALBP-HRC with the cycle time and the number of operators (humans and robots) as the primary and secondary objective, respectively. In addition to the traditional ALBP constraints, the human and robot characteristics, in terms of task times, allowing multiple humans and robots at stations, and their joint/collaborative tasks are formulated into a new mixed-integer linear programming (MILP) model. A neighborhood-search simulated annealing (SA) is proposed with customized solution representation and neighborhood search operators designed to fit into the problem characteristics. Furthermore, the proposed SA features an adaptive neighborhood selection mechanism that enables the SA to utilize its exploration history to dynamically choose appropriate neighborhood operators as the search evolves. The proposed MILP and SA are implemented on real cases taken from the automotive industry where stations are designed for HRC. The computational results over different problems show that the adaptive SA produces promising solutions compared to the MILP and other swarm intelligence algorithms, namely genetic algorithm, particle swarm optimization, and artificial bee colony. The comparisons of human/robot versus HRC settings in the case study indicate significant improvement in the productivity of the assembly line when multiple humans and robots with collaborative tasks are permissible at stations. Access publication.