There are several ML algorithms commonly used in practice and supported by all big Tech platforms. Although there are no many published ML cases in supply chain management, there are potentials for ML in following areas:
- Improve quality of data: It has been constant struggling to maintain accuracy of the data critical to supply chain management. Firstly, the master data like lead time, safety stock etc. for each SKU, all of which need to be updated as business changes. Secondly actual stock aggregated from all transactions, which are done manually or at best, semi-automatically.
- Segmentation analysis: Strategies like sourcing, forecasting, safety stock are based on results of segmentation analysis. It is critical to do the analysis efficiently and effectively. More importantly finding a way to monitor and improve the segments being established.
- Decision-making: We all know limitations with demand forecasting, material planning requirement (MRP), Advanced Planning and Scheduling (APS) and others.
There are challenges to apply known ML algorithms to supply chain management. The use cases show ML applications in supply chain. They can act as reference models for supply chain professionals looking to tackle similar supply chain problems in their own organizations.