A Matheuristic Approach for the Aircraft Final Assembly Line Balancing Problem Considering Learning Curve

In an aircraft final assembly line, the learning effect emerges during the production ramp-up phase. The learning effect results in a decrease in task processing time as production increases, which has a significant impact on the performance of line balancing. The cycle time of the assembly line needs to be adjusted appropriately to guarantee the line's efficiency. This paper addresses an aircraft final assembly line balancing problem with learning curve. The objective is to determine task assignment and cycle time adjustment periods such that the length of the learning stage is minimized. A mixed-integer linear programming formulation injected with an extended DeJong curve is developed to model the problem. A matheuristic approach based on a variable neighborhood search algorithm and a dynamic programming (DP) method is designed to solve the problem. The computational results reveal the superiority of our approach and that the DP obtains the best cycle time adjustment strategy quickly.