Estimation of Effort in Software Projects using Genetic Programming

Abstract

Software effort estimation refers to the estimation of effort that is required in given software project. It starts at the proposal stage and can sometimes continue till the last stages of a software project. Projects normally have a budget, and continual cost estimation is necessary to ensure that spending is in line with the budget. There is need of finding a good model which can establish an accurate relationship between the software a project and cost drivers. It is important for project managers and the researchers working in the domain to explore, analyze and understand the strengths and weaknesses of various software cost estimation methods. This paper focusses on using Genetic Programming for software effort estimation. The implementation involves evolution of individuals for obtaining best results over several generations. Metrics are chosen to evaluate the model based on the literature survey. Standard software engineering datasets are used in this project so that suitability and possible relations that arise could be realized. K-fold validation is used to sum up with more reliable values of evaluation. The design, implementation and result presentation are completed successfully and recorded clearly.