Maximum Likelihood Estimates for the Discrete Application of the AMSAA Growth Model.
Abstract
The U.S. Army Material Systems Analysis Activity (AMSAA) reliability growth model was developed under the leadership of Dr. Larry Crow during the mid to late 1970's. The resultant model was incorporated into Military Handbook 189, Reliability Growth Management, where it is fully explained. The model as developed by Dr. Crow is intended for use with continuous data. However, Dr. Crow later developed an as yet unpublished technique for applying the model to discrete data. Briefly, this technique assigns a sequence number to each trial of the test and uses these sequence numbers as the measures of time in the model. The growth parameters are estimated from the actual test data (the number of test intervals, the number of trials in each interval, and the number of failures in each interval) by the maximum likelihood method. If the resultant likelihood equation had a true maximum, then any of several search algorithms could be applied to find it. However, as will be shown later, the likelihood equation can, in certain regions, be unbounded. The problem only becomes solvable if the maximum over a restricted region is sought. The purpose of this report is to present a solution to the problem and a computer program that implements that solution.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1987
- Accession Number
- ADA186921
Entities
People
- William P. Clay
Organizations
- United States Army Materiel Systems Analysis Activity