Task allocation of human-robot collaborative assembly line considering assembly complexity and workload balance

Human-robot collaboration is increasingly utilised in assembly lines, where task allocation is critical. To address the task allocation problem, this paper first evaluates each assembly task using the indicator of automation potential to determine if a collaborative robot can complete it. The method for evaluating assembly complexity and workload is then introduced, which determines the assembly complexity of each task for both robots and workers, as well as the workload for workers. Based on the above indicators, a new task allocation optimisation model for the human-robot collaborative assembly line is established with the objectives of minimising the cycle time, minimising the workload variance between different workstations, and the assembly complexity per unit product. An improved multi-objective migratory bird optimisation algorithm with fast non-dominated sorting is developed to solve the mathematical model of this task allocation. Finally, the proposed method is applied to an assembly line in a real enterprise. The results of algorithm comparisons show that the proposed algorithm is effective, and some managerial insights are also derived from the experimental tests. The result shows that the study effectively reduces product assembly complexity and balances workers' workload across stations while maintaining assembly efficiency.