2335. The costs of human-centric planning – a constraint programming approach for the AMR fleet in assisted order picking
Invited abstract in session TD-10: Order Picking, stream Supply Chain Management and Production.
Thursday, 14:30-16:00Room: H16
Authors (first author is the speaker)
| 1. | Minqi Zhang
|
| Wirtschaftswissenschaften, Saarland University | |
| 2. | Eric Grosse
|
| Wirtschaftswissenschaften, Saarland University | |
| 3. | Simon Emde
|
| Friedrich-Schiller-Universität Jena |
Abstract
Recent research has increasingly emphasized the integration of ergonomic considerations into industrial operational planning models. Despite this progress, many models continue to simplify realistic resource constraints to reduce computational complexity. This can significantly limit their practical applicability, particularly in environments involving human-robot collaboration. This study addresses this limitation through the scenario of AMR-assisted order picking (AOP). Human workers are responsible for item picking, while AMRs handle transport tasks. Incorporating human fatigue into the planning model has the potential to reduce long-term injury risks. Meanwhile, it raises concerns about both operational costs, such AMR-fleet size, and computational costs, such as model complexity. We propose two constraint programming (CP) models for AMR fleet planning, situated within a broader human-centric framework. The main modeling assumptions include the following: (i) orders, composed of multiple individual tasks, are assigned to AMRs as indivisible units; (ii) each AMR tour, which starts and ends at a central depot, is subject to a limited batch size; (iii) synchronization with human workflows is achieved by imposing strict arrival time windows at each task location, explicitly avoiding waiting time for humans. The planning problem, thus, inherits the computational challenges of typical task allocation and routing problems in logistics and warehousing. Computational experiments using IBM CP Optimizer and Google OR-Tools were conducted to evaluate the feasibility and efficiency of the proposed CP models. Additionally, we discuss how these models can be embedded into a more comprehensive human-centric AOP planning framework through logic-based Benders decomposition.
Keywords
- Constraint Programming
- Computational Experiments
- Warehouse Design, Planning, and Control
Status: accepted
Back to the list of papers