Team-RUP: An Agent-Based Simulation Study of Team Behavior In Software Development Organizations

Abstract

Software production methods are enacted via the interactions of software teams that cooperate to build software. Therefore, organizational culture has a significant effect on project coordination. Yet, this is not reflected in current software process simulation efforts. This thesis introduces a new simulation model development framework, called Team-RUP, to facilitate exploration of the effects of team behavior on the efficiency and effectiveness of software development organizations that pursue incremental and iterative processes such as the Rational Unified Process (RUP). Team-RUP organizes teams according to the degree of autonomy in collaboration and the degree of concurrency in coordination, resulting in four distinct team archetypes: Autonomous, Agile, Concurrent, and Synchronized. Each team archetype embodies a unique combination of autonomy and concurrency levels, highly reflective of modern organizational paradigms. Using Team-RUP, we explore the effectiveness and efficiency of team archetypes under various v levels of task complexity and stability (i.e., internal and external turbulence), as well as team size and workload factors. The conclusions of the simulation study support the claim that process agility is a valid and useful counterbalance to the inevitable change involved in most real-world software projects. In particular, small organizations should consider adopting a software process that encourages agile behavior. If greater independence among teams is necessitated by a particular project, a large organization will perform significantly better than a smaller one.