Estimating Cost Uncertainty Using Monte Carlo Techniques
Part of a study to improve techniques for estimating costs of future weapon systems. The Memorandum presents a technique for expressing such cost estimates as probability distributions that reflect the uncertainty of the estimate. This information is shown to be relevant to the decisionmaking process. The study depicts the relationship between the sources of uncertainty and system cost estimates as an input-output model and, within this framework, a procedure is developed to estimate probability distributions for each of the input uncertainties. From the input distributions, a Monte Carlo procedure is used to generate a series of cost estimates. A frequency distribution and common statistical measures are then prepared from the output estimates to determine the nature and magnitude of the system cost uncertainty. A case study involving the cost estimate of a hypothetical aircraft system with air-to-surface missiles is presented.