The Compare of Solo Programming and Pair Programming Strategies in a Scrum Team: A Multi-agent Simulation

Abstract

Scrum is a popular software development process that addressed the complex software delivery through multi-sprint. However, the scrum team working strategy remain undefined, such as how-to delivery the software at each sprint by optimized task allocation for various team composition is a complex problem. This paper proposed a multi-agent based simulation methods to model scrum team and the proposed innovative team working strategy, in order to verify the performance of the team under different working strategies. A solo strategy and pair programming strategies for the scrum team are designed. Criteria such as completion time, effort time and idle time are considered into the evaluation of the designed team strategies. The testing experiment shows that the adoption of pair programming to scrum team can help the team to improve its performance under specific context.