A Discrete Event Simulation Model to Evaluate Changes to a Software Project Delivery Process

Process simulation modeling has been successfully applied to software development processes to support management decisions on issues related to process management and process improvement. In this work, we develop a discrete event simulation model to study an IT department's project delivery process and estimate the effect on its performance after applying a change to it. We focus on three important performance indicators of project management processes: cost, time and quality. The model uses statistical data analysis techniques and concepts of earned value management to simulate the performance of process activities, model the linkages between them, as well as capture the uncertainty that exists in real-life software processes. Design of experiments is used to organize the experiments and evaluate the results. Finally, two use case scenarios are presented to demonstrate how the simulation model can support decision making during a process improvement initiative. The project delivery process of a major mobile telecommunications provider was used as a template to design the simulation, calibrate parameters, and test the validity of the model.