A Decision Theoretic Approach to the PERT Subjective Estimation Problem.
Abstract
In the PERT subjective estimation problem the objective is to obtain subjective estimates of the unknown parameters of an activity time distribution. The decision theoretic approach suggested in this thesis is based on a single 'most likely' prediction, and it transfers the burden of subjective estimation from the 'expert' to the decisionmaker. No assumption is made regarding the form of the activity time distribution; instead attention is focused on the distribution of the 'expert's' prediction given the unknown parameters of the activity time distribution. Bayes estimates for the mean and standard deviation are investigated for a generalized absolute deviation loss function, and examples assuming a uniform distribution for the 'expert's' prediction are presented. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1970
- Accession Number
- AD0874192
Entities
People
- Charles W. Albo
Organizations
- Air Force Institute of Technology