Decisions related to warehousing are critical in modern supply chains. Like many other supply chain activities, warehousing is undergoing a wave of innovations driven by machine learning, and the body of research in this field is rapidly expanding. By means of a systematic literature review, this work aspires to elucidate the role that machine learning can play in warehousing decision making. To this end, this work analyzes extant research using the DECAS framework, which integrates the classical decision theory elements with elements specifically related to data-driven decision making. This approach introduces the decision-making perspective into the research on this topic, which is still often focused on technical aspects rather than managerial ones. The analysis highlights some interesting results. First, machine learning has been applied to several warehousing decisions, including, but not limited to, item allocation, task allocation, batching, and inventory management. However, most of the applications are based on simulated data, showing that data availability remains a challenge hindering the evaluation of machine learning's effectiveness in real-world scenarios. Moreover, the majority of applications focus on structured decision-making, a domain well-suited to high levels of automation by machines or algorithms. As a result, operative-level decisions are the most commonly explored, while strategic-level decisions which often require hybrid human-machine collaboration-remain largely unaddressed. From an academic point of view, the decision-making perspective introduced in this work can foster further research on the application of machine learning to warehousing decisions, but it can also be replicated in other logistics and supply chain-related contexts. In particular, this work can lay the foundation for research involving real-world applications. From a managerial perspective, this work can help practitioners understand what techniques are more suited for specific warehousing decisions and how much they have already been investigated and tested.
The role of machine learning in warehousing decision making: a systematic literature review
Mascheroni, Matteo;Moretti, Emilio;Garrido, Juan Daniel Rodriguez;Tappia, Elena
2025-01-01
Abstract
Decisions related to warehousing are critical in modern supply chains. Like many other supply chain activities, warehousing is undergoing a wave of innovations driven by machine learning, and the body of research in this field is rapidly expanding. By means of a systematic literature review, this work aspires to elucidate the role that machine learning can play in warehousing decision making. To this end, this work analyzes extant research using the DECAS framework, which integrates the classical decision theory elements with elements specifically related to data-driven decision making. This approach introduces the decision-making perspective into the research on this topic, which is still often focused on technical aspects rather than managerial ones. The analysis highlights some interesting results. First, machine learning has been applied to several warehousing decisions, including, but not limited to, item allocation, task allocation, batching, and inventory management. However, most of the applications are based on simulated data, showing that data availability remains a challenge hindering the evaluation of machine learning's effectiveness in real-world scenarios. Moreover, the majority of applications focus on structured decision-making, a domain well-suited to high levels of automation by machines or algorithms. As a result, operative-level decisions are the most commonly explored, while strategic-level decisions which often require hybrid human-machine collaboration-remain largely unaddressed. From an academic point of view, the decision-making perspective introduced in this work can foster further research on the application of machine learning to warehousing decisions, but it can also be replicated in other logistics and supply chain-related contexts. In particular, this work can lay the foundation for research involving real-world applications. From a managerial perspective, this work can help practitioners understand what techniques are more suited for specific warehousing decisions and how much they have already been investigated and tested.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
