Agent-based modelling for human–robot collaborative order picking system considering workers’ performance
研究了物流仓库中自动导引车与工人协作的订单拣选系统，通过智能体建模比较不同工人角色对系统性能、AGV利用率和安全性的影响，发现考虑工人疲劳和学习能提升绩效，且最优场景取决于人机比例。
In a logistics warehouse, automated guided vehicles (AGVs) collaborate with humans to meet the growing demand for logistics services. Within a human–AGV collaboration, the flexibility of humans allows them to perform various roles. This study compares the performance of collaborative order picking systems across various human roles, aiming to understand their impact on the order picking system. We propose three order picking scenarios based on different human roles: retrieval task focused, partially carried by human, and multi order picking methods. We implement these dynamic scenarios using agent-based modelling and evaluate the systems in terms of performance, AGV utilisation, and safety, varying the number of human agents. To obtain precise simulation results and improve the matching process between AGV agents and human agents, we incorporate human factors such as fatigue and learning. Results confirmed an improvement in system performance due to the consideration of changes in human performance during the matching process. The optimal performance scenario varied according to the ratio of human agents to AGV agents, emphasising the importance of resource consideration when defining human roles. Additionally, we found that potential safety issues for workers increased in environments with high AGV utilisation.