Robustness in Cost Performance Models.
Abstract
This thesis examined the robustness of methods used to compute estimates at completion of contract costs. The four techniques examined were a subset of those used by the Cost Performance Report Analysis computer program to derive estimates at completion. Monte Carlo simulation methods were used to generate the required data values for 18 simulated populations representing a variety of possible cost and schedule overrun scenarios. A statistical evaluation of the robustness of the four estimate at completion techniques was performed using samples from each of the 18 simulated populations. The results of this analysis indicate that, for practical purposes, none of the selected estimate at completion techniques are robust for the cost and schedule overrun scenarios that were investigated. The tendency of each of the estimate at completion models was to underestimate the final contract cost. The estimate at completion models incorporating mathematical smoothing techniques (i.e., weightings or moving averages) provided relatively stable results across all of the populations; however, they were unable to consistently predict a contract's completion cost within 30 percent of the actual value.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1987
- Accession Number
- ADA187183
Entities
People
- Darren E. Morgan
Organizations
- Air Force Institute of Technology