An advanced cost estimation methodology for engineering systems

A mathematically advanced method for improving the fidelity of cost estimation for an engineering system is presented. In this method historical cost records can be expanded either through the use of local metamodels or by using an engineering build-up model. In either case, the expanded data set is analyzed using principal component analysis (PCA) in order to identify the physical parameters, and the principal components (PCs) which demonstrate the highest correlation to the cost. A set of predictor variables, composed of the physical parameters and of the multipliers of the principal components which demonstrate the highest correlation to the cost, is developed. This new set of predictor variables is regressed, using the Kriging method, thus creating a cost estimation model with a high level of predictive capability and fidelity. The new methodology is used for analyzing a set of cost data available in the literature, and the new cost model is compared to results from a neural network based analysis and to a cost regression model. Further, a case study addressing the fabrication of a submarine pressure hull is developed in order to illustrate the new method. The results from the final regression model are presented and compared to results from other cost regression methods. The technical characteristics of the new novel general method are presented and discussed. © 2011 Wiley Periodicals, Inc. Syst Eng © 2012 Wiley Periodicals, Inc.