Machine Learning based Backlog Prioritization Techniques in Distributed Agile Software Development

Abstract

Distributed Agile Software Development (DASD) [5], is the main software development trend in today’s competitive world. In this environment, team members dispersed at different locations globally, work on a common Agile development framework. Agile based development techniques make use of Sprint cycles [6], which is a two week development cycle. At the end of each sprint shippable deliverables are delivered to customers for acceptance and feedback. Based on these feedbacks and relevant inputs from the customers, a list of new features, amendments to present features, bug fixtures and other tasks are identified. These tasks are required to be delivered by teams to achieve outcomes as per the desire of customer and they are referred as product backlogs. These backlogs need to be arranged into an order of execution for completion and timely delivery of the product. This ordering is termed as backlog prioritization. In this paper we have done survey of different backlog prioritization techniques used by the researchers for Agile software development. Further, we have proposed a model for backlog prioritization based on machine learning mechanisms.