The effects of team backlog dependencies on agile multiteam systems: A graph theoretical approach

Abstract

In agile software development, the prioritization of backlog items is the mission-critical responsibility of the product owners in order to maximize the customer value created by development teams. However, in the reality of large-scale development, the degree of freedom for such a prioritization is substantially restricted by various types of interdependencies between backlog items. In this work, we show, using a graph theoretical approach, the relation between the degree of freedom for prioritization and the occurrence of dependencies. To the best of our knowledge, the breadth and depth of such consequences has never been modeled or investigated up until now. Based on our results, we derive implications for real-world large-scale software development in agile environments.