Software cost estimation using homeostasis mutation based differential evolution

Abstract

The main concern in the field of software development is estimation of the cost of software at its initial phase of development. The cost estimation usually depends upon the size of the project, which may use lines of code or function points as metrics. In COCOMO, for the accuracy of the cost estimation, cost factors need to be formulated in the individual development environment. In this paper, some new mutation strategies are proposed to improve the accuracy of cost estimation by modifying parameters of COCOMO using Homeostasis mutation based differential evolution(HMBDE). The proposed method adds one more vector named as Homeostasis mutation vector in the existing mutation vector to provide more bandwidth for selecting effective mutant solutions providing a wide search space for probable solution. The proposed approach provides more accurate solutions to guide the evolution. Performance of proposed algorithm is compared with software cost estimation models. The result verifies that our proposed HMBDE performs better than COCOMO based DE and PSO algorithm and other soft computing models.