An Analysis of the Demand Distributions for Selected Inertial Measurement Units in the Air Force Inventory
Abstract
Within the Air Force Logistics Command's Recoverable Item Requirements Computation System (D041), a simple Poisson demand distribution is assumed in order to forecast requirements and, thereby, to establish inventory levels for recoverable items. This study is pursuant to the question of the validity of this assumption. Detailed analysis of failure data for the FLIP, LN- 15, and KT-73 inertial measurement units revealed that the demand distribution for these items was not Poisson, but more generally, negative binomial. Use of a negative binomial demand distribution in the Variable Safety Level (VSL) portion of Do41, however, requires that the steady state probabilities for that demand distribution be computed. Thus far, this computation has proven to be intractable. Further research uncovered the possibility that a compounded form of the Poisson distribution might also describe the failure data, but research in this area was deferred because SIMFIT, the data analysis program used in the study, did not include subroutines to examine this sort of distribution.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1978
- Accession Number
- ADA059287
Entities
People
- David R Johnson
- Judith M. Mccoy
Organizations
- Air Force Institute of Technology