Mixed case palletizing is a real-world application variant of the 3D Bin packing problem, where the goal is to stack mixed size cases onto a pallet in such a way it optimizes the space utilization i.e., stack as many cases onto a minimum amount of pallets. Although widespread applications in logistics exist, it remains a challenging problem, and stacking of mixed size products still largely involves manual labour.
Mixed case palletizing is subject to additional challenges as it not only requires robotics to grab the oncoming item i.e., determine the position of the product and robot kinematics to actually pick it up, but also to determine the size and optimal placement position on the pallet.3 In recent years, some research has utilized Deep Reinforcement Learning, where robotic agents aim to learn an optimal placement position by stacking many pallets in simulation. Some works apply there learned model in real-world environments.4
Kumar, Shrawan; Mitalb, Anil; Garanda, Doug; Persada, A. (1993). "Operator stress in palletizing tasks with restricted access and headroom". International Journal of Industrial Ergonomics. 12 (1): 153–162. doi:10.1016/0169-8141(93)90046-G. /wiki/Doi_(identifier) ↩
Brunelli, Richard (Fall 2004). "Packaging Basics: Beasts of Burden". PMTdirect.com. Packaging Machinery Manufacturers Institute. Archived from the original on October 9, 2011. https://web.archive.org/web/20111009161629/http://www.pmtdirect.com/website/article.asp?id=1806 ↩
Viroteq. "Building an Automated Palletizer for Mixed Parcel Stacking". Viroteq.ai. Retrieved 4 December 2023. https://www.viroteq.ai/automated-palletizer-mixed-parcel-stacking/ ↩
Zhao, Hang; She, Qijin; Zhu, Chenyang; Yang, Yin; Xu, Kai (2022). "Learning Efficient Online 3D Bin Packing on Packing Configuration Trees". Thirty-Third Conference on Innovative Applications of Artificial Intelligence (IAAI): 741–749. S2CID 251649027. /wiki/S2CID_(identifier) ↩