Thanks to a Phase I Small Business Innovation Research (SBIR) grant from the National Science Foundation (NSF), demand forecasting solution vendor Smart Software will investigate new statistical methods to forecast intermittent demand. The research seeks to extend demand forecasting beyond individual products and parts, identifying and interpreting interactions across clusters of items whose demands fluctuate together.
The new research will build upon the software vendor’s existing method for forecasting slow-moving or intermittent demand, developed with the support of a previous NSF grant. The current method evaluates historical demand for each item and establishes the optimum level of inventory that will be required to achieve service level objectives.
The new forecasting capabilities are expected to include:
· A more dynamic statistical model of parts to enable forecasts that better reflect a variety of external factors including part usage by itself or in combination with other products, as well as the impact of macroeconomic and environmental factors;
· A dynamic model of item usage, enabling planners to develop functional maps of the interrelationships of large numbers of parts. Knowing which parts have demands that co-vary can be useful in at least two ways. First, item managers can be assigned to work with coherent clusters rather than arbitrary collections of miscellaneous parts, and second, parts can be co-located in warehouses for more efficient storage and retrieval.
· Improved forecasts of "aggregates" where intermittent demand is present, such as all items in a product line, or all items at a particular warehouse. This would be useful for raw materials purchasing, as well as for financial planning when parts are a source of revenue.
“Any organization that builds or supports capital equipment experiences intermittent demand for some portion of its inventory,” said Nelson Hartunian, president of Smart Software. “This grant is a terrific opportunity to impact one of the biggest forecasting challenges facing these organizations—accurately forecasting parts and optimizing inventories. Ultimately, the goal is to have the right part at the right place at the right time. The research we are undertaking will make this goal more achievable.”
