The proposal addresses development of a tool to improve software cost and schedule estimating accuracy. After collecting and building a preliminary database of historical software cost, technical, and programmatic data, we will explore application of innovative methods and techniques to use in the software estimating tool. The anticipated results of this research include: 1) an increase in the accuracy of software development and maintenance cost and schedule estimates; 2) a more realistic and defensible quantification of uncertainty (estimating error range); and 3) a reduction in the effort required to generate an estimate and calibrate the estimating tool.
Benefits: Our proposed software cost-estimating tool will be tailored to both the unique requirements of Department of Defense (DOD) organizations and to the unique needs of industry, including defense and other commercial concerns. Because the model will provide a more accurate and defensible cost estimate, we believe we will achieve substantial market penetration within DoD. Consequently, defense firms developing and maintaining software would then have a strong incentive to also purchase and use this model in order to: 1) determine what the governments estimate of their project will be and 2) produce an estimate that the government is more likely to consider reasonable. Because of its proposed built-in capability to tailor itself to the salient attributes of a specific software organization, the model can also be rapidly generalized for use by other commercial firms.
Keywords: software, estimating, analysis, prediction, development, maintenance, cost, schedule
