Improved Methodology for Developing Cost Uncertainty Models for Naval Vessels

Abstract

The purpose of this paper is to analyze the probabilistic cost model currently in use by NAVSEA 05C to predict cost uncertainty in naval vessel construction and to develop a better method of predicting the ultimate cost risk. The data used to develop the improved approach is collected from analysis of the CG(X) class ship by NAVSEA 05C. The NAVSEA 05C cost risk factors are reviewed and analyzed to determine if different factors are better cost predictors. The impact of data elicitation, the Money Allocated Is Money Spent (MAIMS) principle, and correlation effects are incorporated into the research and analysis of this paper. Data quality is directly affected by data elicitation methods and influences the choice of probability distribution used to give the best predictor of cost risk. MAIMS and correlation effects are shown to make a significant impact to the overall cost model. Program managers and analysts can readily implement the enhanced models using commercial Excel add-ins, such as Crystal Ball or @Risk and integrate them into their current cost risk analysis and management practices to better mitigate risk and control project cost.