Network-Embedding Based Storage Location Assignment in Mobile Rack Warehouse

Abstract

As mobile rack warehouses become more and more popular in e-commerce era, traditional storage location assignment strategy which optimize the space, retrieval speed, utilization ratio is no longer suitable for such situation. Current mobile rack warehouse often using random strategy to put goods onto racks. However, this strategy doesn’t consider the relationships between goods, which are implied in order information. In this paper, a Network-Embedding based method is proposed to cluster goods into different groups, which helps to create storage location assignment strategy. First, we build the relationship network between goods based on the history orders data. Then, we train the goods representations through the network embedding model. At last, we find the strong-related goods by K-means algorithm, and put them onto the same rack. The experimental results show the method we proposed is more efficient than random strategy.