Evaluation of Variance Approximations and Demand Forecasting Techniques
Abstract
ALRAND Working Memorandum 365 of 26 September 1980 recommended two potential improvements to the current UICP (Uniform Inventory Control Program) demand forecasting techniques: (1) a more direct approximation for the variance of quarterly demand; and (2) adaptive smoothing to forecast demand. This study used the 5A (Aviation Afloat and Ashore Allowance Analyzer) wholesale inventory simulator and actual demand observations to compare the suggested alternatives to the current method. The following criteria were used in the comparison: inventory investment, performance, workload, demand forecast accuracy and the required computer time. However, the primary criterion used was the change in performance per dollar invested. The study showed that (1) the more direct approximation of standard deviation of demand is not an improvement; and (2) adaptive smoothing with filtering should be considered for implementation.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 31, 1981
- Accession Number
- ADA111220
Entities
People
- Kim L. Hoff