Authors:
Dolev Starr, Luis Pujol
MIT Supply Chain Management Program (SCM)
Summary:
Our research aims to improve the forecasting accuracy of Naval Aviation MRO (Maintenance, Repair, and Overhaul) consumables for the Defense Logistics Agency (DLA) by applying a “cluster-then-forecast” approach. This methodology involves clustering SKUs into categories based on volume, volatility, and intermittency, and then applying distinct categories of forecasting models, such as statistical, Machine Learning, and Deep Learning. Lastly, we will compare each model for each cluster using both classical and contemporary forecasting metrics, with the goal of recommending which model to use for each cluster.
Leave a Reply