A Novel Technique of Software Cost Estimation Using Flower Pollination Algorithm

Estimating Cost of Software (ECS) is a very essential aspect of a software development life cycle. Accurate estimations are needed in terms of person month and development time for software projects. Several project estimation techniques have been developed such as parametric and non-parametric. Estimation by parametric is one of the convenient methods in the field of software engineering. However, parametric approaches used in the development of software cost estimation is not able to handle the definite data in an explicit and accurate way. Different computational intelligence algorithms like machine learning, evolutionary algorithms, and swarm intelligence algorithms have been applied to optimize the parameters of various ECS model. But however, accurate cost estimation is still a big issue in ECS. In this work, a Flower Pollination Algorithm (FPA) is proposed to optimize the parameters of a Constructive Cost Model II (COCOMO-II) via using a standard Turkish industry dataset. Experimental results demonstrate that the proposed algorithm gives a better estimation as compared to existing approaches such as the Bat algorithm and original COCOMO-II in terms of Manhattan distance (MD) and mean magnitude of relative errors (MMRE).