An Analysis of Models for Forecasting Repairable Carcass Returns.
Abstract
This thesis evaluates techniques for forecasting the return of failed repairable spare parts (known as carcasses) within the U. S. Navy supply system by comparing the model currently implemented i the Uniform Automated Data Processing System Inventory Control Point (UICP) program with several alternative forecasting models to determine if an improvement can be achieved in forecasting effectiveness. The current model uses an exponential smoothing procedure and applied several filtering processes to determine the appropriate smoothing constant value. The alternative models employ forecasting techniques such as moving average, moving least squares, adaptive response rate, and regression analysis. Each model is then synthesized with actual U. S. Navy supply system data and its performance measured by a set of evaluation criteria. The results indicate that the current UICP forecasting model cannot be improved substantially and that a filtering process is critical to the performance of any model applied to real world data.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1982
- Accession Number
- ADA124606
Entities
People
- Douglas Martin Hartman
Organizations
- Naval Postgraduate School