Graph Neural Networks and Deep Reinforcement Learning for Warehouse Order Picking and Representation Learning

Order picking is one of the most labor-intensive warehouse operations, and improving routing efficiency remains an important challenge. While deep reinforcement learning (DRL) has shown promise in complex optimization problems, its application to warehouse order picking is still limited, and graph-based representation learning using graph neural networks (GNNs) in this context remain largely unexplored. This paper proposes a GNN-based DRL method that models warehouse layouts as graphs to optimize order-picking paths while simultaneously learning graph-based structural embeddings of storage locations. The approach is evaluated in simulated warehouse environments of different scales and benchmarked against classical heuristics, including the Lin–Kernighan algorithm. The results show that the proposed GNN–DRL approach consistently achieves shorter travel distances than traditional methods, particularly for larger orders, and remains effective across different warehouse layouts when fine-tuned. Moreover, the learned node embeddings capture meaningful structural properties of warehouse layouts and adapt to different operational contexts, highlighting the potential of integrating GNNs and DRL as a flexible foundation for advanced warehouse optimization.