An Investigation of the Accuracy of Heuristic Methods for Cost Uncertainty Analysis
Abstract
Cost uncertainty analysis has received a great deal of attention over the last several years. The purpose of a cost uncertainty analysis is to identify the cost and schedule implications associated with program uncertainties. Common methods for uncertainty analysis characterize the possible cost and schedule outcomes of a project using a probability density function (pdf). Heuristic methods have been proposed for uncertainty analysis that assume the shape of the total cost pdf is either normally or lognormally distributed. While experienced analysts feel these distributions provide reasonable approximations, little evidence exists to either confirm or refute these presumptions. An experiment is conducted in which number of cost elements, the degree of skewness of the cost elements, and the degree of correlation between cost elements are varied systematically. The resulting total cost pdfs are compared to the heuristic distributions using goodness of fit tests. The results show that the normal distribution provides an excellent approximation for the simulated distributions. Guidelines are offered that help the cost analyst determine whether these heuristics ought to be applied in a cost uncertainty analysis. Cost analysis, Cost uncertainty analysis
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 26, 1994
- Accession Number
- ADA285241
Entities
People
- Kevin P. Grant
- Wendell P. Simpson Iii
Organizations
- Air Force Institute of Technology