Introduction to DLA

The Defense Logistics Agency (DLA) is the Department of Defense’s (DoD) largest combat logistics support agency. DLA jointly provides logistic support to the Army, Navy, Air Force, Marine Corps, other federal agencies, as well as combined and allied forces with logistics, acquisition, and technical services. DLA sources sustenance, fuel and energy, uniforms, medical supplies, and construction and barrier equipment to our military, helping to maintain their ability to operate. DLA also supplies military spare parts, manages restoration of military equipment, while additionally providing catalogs and production services. DLA manages almost 5 million items, fills more than 131,000 requisitions per day, and issues 10,000 contract actions per day.

What is the purpose of the study?

The purpose of this study was to assess DLA’s current methods and policies, identify necessary improvements, and provide recommendations to increase the availability of items for customers. This initiative was scoped to an analysis solely on AAC D and RMC R NIIN’s (National Item Identification Numbers) current inventory processes. Since AAC D and RMC R NIINs have such high demand, improvements to these items’ current inventory allocation and forecast will have the greatest effect on DLA’s inventory management.

What did we do?

Our team addressed the problem using a series of 5 separate efforts. First we conducted a literature review to identify additional inventory metrics,  determine best practices for inventory management, understand the application of Coefficient of Variation (CoV) when determining if a NIIN can be forecasted, and refine appropriate methods to use for cluster analysis. Next, we conducted data exploration - specifically relating to Materiel Availability. We then conducted Cluster Analysis on the data and later on the results to attempt to identify specific clusters within the NIINs. We built a stochastic model in Matlab which searches the trade space of the problem to determine the optimal combination of ordering policies for each NIIN. Lastly, we ran the model and performed analysis on the results.