In the early stages of a software development life cycle, effort estimation plays a critical role in helping project managers predict the demands with respect to the budgeting, scheduling, and the allocation of resources. In this situation, the ideal estimation calculation should provide an approximate value figure, which will consist of a base estimation value plus a contingency allowance value, which will cover the risks and assumptions necessary for particular estimation calculations.
However, most software effort estimation methodologies, which include the COCOMO model, provide a fixed effort estimate value instead of an approximate value, and consequently the existing effort estimation approach has failed to become a trusted reference for project manager due to the problem in estimation accuracy.
This paper introduces the Fuzzy-ExCOM Model, the Software Project Risk Assessment and Effort Contingency Model based on a COCOMO cost factors, which provides the project risk identification and contingency allowance to complement the effort estimation value based on identified project risk and software size. The proposed model also integrates the effort estimation and risk assessment activities because these activities are integral parts of the initial software project planning phase and the accuracy of the effort estimates depends heavily on the nature and level of the risks that are inherent in the software project.
A validation of this model using a project data sets shows that the new model provides a higher level of effort prediction performance compared to the existing COCOMO-II effort estimation approach.
Available at: http://works.bepress.com/luiz_fernando_capretz/7/