Optimization of Worker Assignment for a Manual Assembly Production Lines using Reinforcement Learning

In the technological advancement known as Industry 4.0, the strategic implementation of flow production lines is essential for mass customization. As manual assembly production lines rely on the high flexibility and adaptability of human workers, it is important to consider human-centric, a key value of Industry 5.0, in addition to improving production efficiency. Worker assignment in manual assembly production lines is crucial, as it directly influences both the efficiency of the line and the workload of the workers. However, it is difficult to derive the optimal worker assignment by considering a number of complex factors. Therefore, we propose a data-driven worker assignment methodology using reinforcement learning for multi-objective optimization in manual assembly production lines. A case study is conducted in a laboratory environment that mimics a manufacturing site, and the proposed methodology is validated by comparing it with existing worker assignment methods. It is expected that future research on data-driven multi-objective optimization problems for manual assembly production lines will be diversified based on this paper.