| Current space project cost models predict cost using regression equations that typically relate the cost of projects to technical variables such as Weight, Power, Thrust, Data Rate, and Pointing Accuracy. While improvements in space project parametric cost modeling have been made since its advent in the 1960s, significant statistical error persists. This remains true even when a rich set of technical independent variables is used. Anecdotal evidence suggests that part of this variability may be due to the inability of current cost models to capture the impacts of the non-technical considerations required to develop space systems. A literature review developed a number of such non-technical variables. These include, as examples, the experience of the team implementing the project; the extent of up front formulation of the project prior to full-scale development; the degree that requirements are solidified and maintained after project start; and the availability of timely funding. This research investigates the feasibility of introducing such non-technical variables into the cost models, and then analyzes the improvement in the explanatory power from these additional variables. The research involved constructing a data base of historical projects containing both technical and non-technical characteristics. Subsequently, an extensive multivariable regression analysis was performed to develop three new cost models: one with technical variables only, as a control (which will be referred to as Model I), one with non-technical variables only (Model II) and one with both technical and non-technical variables (Model III). The three models were compared to assess improvement in the predictive power of the models containing non-technical variables, as compared to the more classic approach of focusing mainly on technical variables. An additional analytical approach was performed to introduce a final non-technical variable, Team Experience, with revised versions of Model II and Model III provided which include this variable.;The outcome of the analysis did verify that several non-technical variables can be successfully introduced into space project cost models with acceptable regression statistics. Furthermore, the prediction capability of the models is improved by the inclusion of non-technical variables. In fact, the research found that a successful model can be constructed using non-technical variables only with no technical variables. However, the single best model developed included a mix of both technical and non-technical variables. |
