Improve the Accuracy of Software Project Effort and Cost Estimates in COCOMO II Using GWO

Most of the software industry uses the Constructive Cost Model to be able to estimate the effort and cost of software projects, such as COCOMO II which has a dependence on cost drivers. Value of cost driver can affect the accuracy of the project effort and cost estimate. However, the COCOMO II accuracy value is considered to be less accurate because there is still a large difference between the actual project effort and the cost estimated value. To improve the accuracy of COCOMO II, Grey Wolf Optimization (GWO) method is used which is based on the behavior of wolves in catching prey. In this study, COCOMO II GWO is used to obtain a higher and more accurate level of estimation accuracy and reduce the total error value or Mean Magnitude Relative Error (MMRE) of software projects. From the test result when compared to MMRE produced by previous study COCOMO II BCO (Bee Colony Optimization) was 12.92%. Meanwhile, MMRE by proposed method COCOMO II GWO IS 1.731%. It means the proposed method can reduce the error value in MMRE by 11.19%.