Quantitative Estimation of Cost Drivers for Intermediate COCOMO towards Traditional and Cloud Based Software Development

Abstract

Software project estimation is the process of analyzing the resource requirements for the given time duration of product development. Cost estimation models are used for calculating the associated amount required for developing the stakeholder's requirement within the defined time boundaries. Among several models available for the cost estimation of software projects, COCOMO is one of the well-known models which serve the field most. Resources applied for the given time will generate the rough estimates, but for more accurate values, various factors are analyzed. These factors are termed as cost drivers. Software estimation using COCOMO is performed by selecting values of cost drivers on a predefined scale. This approach solely depends on experience of a software analyst. However, there is a lack of a systematic approach available for the selection of values of these cost drivers. Our work suggests the quantification of cost drivers for intermediate COCOMO. Quantification will implicitly fetch the values from the system and its environment which reduces the manual selection of ranges of scaling factors. Hence the systems cost will be generated directly without analyst and selector logic. Finally, if the selection of correct scaling is performed, then the calculation of cost will definitely get improved. An experimental analysis is performed between the above suggested model and the Intermediate COCOMO. The results show that the "COCOMOUP" is performing well under the known conditions and in uncertain requirements conditions, the system is getting better predictions.