Simulink based fuzzified COCOMO

Abstract

Accurate estimation of the cost of a software minimizes the risk for the software development process. Applicability of the different cost estimation models is very crucial as the information required for the implementation of such models is imprecise and vague. In order to increase the accuracy of the model, it is proposed to introduce the concept of fuzzification for calculating the value of effort multiplier corresponding to different of cost drivers used in COCOMO model. The proposed model is designed using the MATLAB and is modeled in SIMULINK. The results are validated using the COCOMO dataset and it has been observed that by fuzzifying the cost drivers the model generated the results more closer to the actual values and thus enhances the accuracy of the estimation process.