NAVSUP to Employ New Wholesale Inventory Optimization Model

Nov. 6, 2019 | By Office of Corporate Communications

Office of Corporate Communications, Naval Supply Systems Command

NAVSUP Weapon Systems Support (WSS) is the Navy’s Program Support Inventory Control Point, and, in coordination with the Office of the Secretary of Defense, has been focusing on how to improve methods for establishing wholesale inventory policy, the backbone of the Navy supply system.  “The new Wholesale Inventory Optimization Model for Enterprise Resource Planning (WIOM-ERP) will enable NAVSUP to make better, more robust, and more defendable inventory policy decisions,” said Lt. Cmdr. Duncan Ellis, NAVSUP Headquarters operations research branch director.

The landscape NAVSUP WSS manages is a $34 billion inventory, processing more than 500,000 equipment and repair parts demands each year for nearly 300 deployment battle ready ships, 92 submarines, and more than 3,700 operational aircraft.  Establishing target stock levels for both ready for issue and not ready for issue items, that meet customer demand, is a large challenge for inventory managers who must consider a multitude of factors. To manage this complex process, NAVSUP uses Navy Enterprise Resource Planning (ERP) supported by optimization models like the legacy Wholesale Inventory Optimization Model (WIOM).

NAVSUP is developing next-generation modeling tools to help compensate for the significantly complicated inventory management problem of the low and often highly intermittent demand of most items that NAVSUP WSS manages, with many items only receiving one or two demands over a five-year planning period.  Recent analyses conducted in coordination with Penn State Applied Research Laboratory, reveal that this item profile means less than 30% of Navy managed items perform well when planned for with traditional forecasting techniques. To plan for this difficult portfolio, the next-generation modeling tools like WIOM-ERP combine simulation and optimization to offer higher performance than the legacy models.  Optimization models are often used in supply planning to set, or prescribe, inventory policy across large populations of items. An optimization model attempts to meet a specified objective while subject to constraints. Models like legacy WIOM can be extraordinarily powerful in prescribing optimal policy; however, they have limitations when used on low or intermittently demanded items.

Simulation models are widely used and accepted across industry, academia, and government with a wide range of applications. Simulations offer enormous utility in understanding the behavior of complex systems where real world testing is either impractical or impossible.

“Simulation, by itself, has limited value for setting enterprise inventory policy, we would quickly find ourselves with an overwhelming amount of model output data,” said Ellis. “However, we can address the challenge by leveraging the capability of optimization models to prescribe policy that is ‘globally optimal’ given an objective, constraints, and a decision space.”

WIOM-ERP is a next generation modeling suite, which leverages the large volumes of historical data collected by our systems and pairs a sophisticated simulation and optimization model in order to establish inventory policy for large populations of items.

The WIOM-ERP software suite combines three main components, a controller, a discrete event simulation (DRPSim), and an optimization model (WIOM-OPT). The coupling of the DRPSim with WIOM-OPT allows for more reliable setting of inventory levels across a wider range of materiel than possible with current tools.  WIOM-ERP is being developed under a technical partnership with the Naval Postgraduate School in Monterey, California.

The controller uses raw data, to include historical requisitions, repairs, and procurements that are combined with item information and variables to create a WIOM input set. Input sets can contain millions of rows of transactional data.  The controller automatically handles data extraction and conducts required validation and restructuring.

DRPSim is a sophisticated event simulation that models the supply system against multiple generated synthetic futures. The simulation models both internal and external events to model what might happen under each of these synthetic futures. The simulation can return both detailed transactional level detail such as inventory levels over time, and summarized metrics such as itemlevel requisition fill rates.

WIOM-OPT is a multi-objective optimization model that seeks planned minimum safety stock level values that are optimal for all items simultaneously considered.  WIOM-OPT establishes weights with the optimization model framework and is designed to give large amounts of flexibility to the user. For example, a user may need to produce recommended levels that minimize customer wait time while constraining inventory below a threshold. This flexible architecture offers a form of future proofing in that the model can flex to meet changing needs or shifts in organizational values.

The WIOM-ERP development team is targeting an aggressive timeline with development and testing complete by the summer of 2019 followed by an initial production implementation scheduled for January 2020.

“WIOM-ERP is the future of wholesale inventory modeling for NAVSUP,” Ellis concluded.

Improved inventory modeling yields a direct benefit to our customers, ensuring NAVSUP investment in wholesale materiel is positioned to provide higher readiness, better responsiveness, and increased trust and confidence.