Document Type : Research Paper

Author

• Amid Khatibi Bardsiri

Department of Computer Engineering, Bardsir Branch, Islamic Azad University, Bardsir, Iran

Abstract

Today, software projects are a major part of information and business technology. Estimating the cost and delivery time of a project is one of the most important aspects of its development process. The results of this estimate are directly related to the failure or success of the whole project. Project anomalies, high diversity, intangibility, and poor standards are obstacles that have hampered various current forecasting models. In this regard, the aim of this paper is to provide an intelligent hybrid model that is able to weigh the project features properly. The main idea of the proposed model is based on mathematical methods, soft computing and using previous estimation methods. After the complete introduction of the new model, its efficiency is evaluated using Desharnais and ISBSG databases. The results indicate higher accuracy and greater flexibility of the proposed model. Improvement rate shows in the various metrics such as MdMRE, Pred(0.25) and MMRE separately, but the average improvement rate is 31%.

Keywords

• Intelligent model
• Analogy based estimation
• Differential evolution algorithm
• Budget and time estimation