Most of software industry are using Constructive Cost Model (COCOMO) to estimating software cost. COCOMO II has a high dependency on cost drivers, and the role of cost driver assessed to improve the reliability of the estimated effort. COCOMO II is important to be emphasized at uncertainty in the level of input which can generates the uncertainty output and it can leads to the error of estimations. To improve the accuracy of COCOMO II, then it used a trapezoidal membership function of fuzzy logic which is represented by cost driver in COCOMO II. COCOMO II used as the basis of research and used as part of experimental dataset research to describe the approach and compare it with the standard version of COCOMO II. It has been explained that the function of trapezoid can performs better that COCOMO II’s function, because it shows a transition which smooth in the interval, and the final results are not far from the actual effort. From this equation, the value of which is smaller than the MRE or MMRE is closer to the real effort. MMRE of the method which proposed can reduce 16.29% from the previous method.
- COCOMO II
- Cost driver
- Fuzzy logic
- Software cost estimation
- Trapezoidal membership function