Context-aware task allocation for distributed agile team

Abstract

The philosophy of Agile software development advocates the spirit of open discussion and coordination among team members to adapt to incremental changes encountered during the process. Based on our observations from 20 agile student development teams over an 8-week study in Beihang University, China, we found that the task allocation strategy as a result of following the Agile process heavily depends on the experience of the users, and cannot be guaranteed to result in efficient utilization of team resources. In this research, we propose a context-aware task allocation decision support system that balances the considerations for quality and timeliness to improve the overall utility derived from an agile software development project.We formulate the agile process as a distributed constraint optimization problem, and propose a technology framework that assesses individual developers' situations based on data collected from a Scrum-based agile process, and helps individual developers make situation-aware decisions on which tasks from the backlog to select in real-time. Preliminary analysis and simulation results show that it can achieve close to optimally efficient utilization of the developers' collective capacity. We plan to build the framework into a computer-supported collaborative development platform and refine the method through more realistic projects.