Probabilistic Estimates with Limited Data.
Abstract
Decision makers require and need more than a point estimate of the cost and schedule for new or on-going programs. Much has been written about how to quantify program risks based upon historical data. However, estimates for new systems frequently depend upon expert opinion from a few knowledgeable persons because an applicable historical data base is not available. The method proposed in this report utilizes a PERT type network, beta distribution parameters for expert opinion inputs, and convolution of activity distributions by simulation. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1980
- Accession Number
- ADA093969
Entities
People
- Conrad W. Faber