Machine Learning Inventory Forecasting

The 147th ATKW is seeking new ways to forecast 9MM ammunition that it maintains at the unit. Currently, each Squadron and personnel require different cycles they must shoot to maintain currency. There is not a forecast methodology presently in place. Due to this, the unit buys more than needed each year and must shoot the oversupply at the end of the fiscal year. Furthermore, the group must store the extra 9mm using valuable space that would be better used for other critical parts. We believe that technology implemented with SparkCognition under the SBIR Phase I topic will solve this forecasting challenge. If the project is successful, the solution Darwin will allow the Squadron to forecast a multitude of other problems. Examples are the weather for pilot scheduling, medical supplies & MQ-9 Reaper parts forecasting, fly away kits, forecasting troop movement resources, etc. With the DOD inventory management systems monitoring millions or parts across the globe given the Air Force a 10% better forecasting tool would save 100 of millions of dollars and increase uptime for the warfighter by having the right parts at the right location at the right time.