Machine Learning for Warehouse Management: A conceptual framework

Warehousing industry is highly affected by trends of globalization and the increasing dynamics of product life cycles, which results in global supply chains. For warehouse management, a huge number of products needs to be delivered on time and at the lowest cost possible in order to provide desired level of customer service. Currently, a high amount of time and energy are spent to gather information about products and customers to be used to decide about the needed warehouse operations and to measure their performance. Therefore, this paper presents an approach to predict future warehouse operations as well as the measurement of performance using an integrated approach. The proposed approach includes four main stages: the identification of warehouse operations and KPIs, then the creation of warehouse operations and KPIs ontology, after the modelization of the ML model and evaluation, and finally the creation and the prediction of future warehouse operations and KPIs. Using machine learning, generic knowledge of warehouse operations and performance measurement could be extracted then used to predict future scenarios.