An Agile Project System Dynamics Simulation Model

This paper compares established Systems Dynamics (SD) models of software projects with models of agile development. A new minimal SIMULINK (TM) agile project model was created and compared to a Waterfall model of a NASA project. Results are presented to enable project managers to predict the performance of future agile project processes. The model includes the time to generate the requirements and a function to provide requirements volatility. These models show that for the same productivity and fraction of errors that are satisfactory, the overall development staff costs are similar for agile and waterfall projects and the undiscovered rework is less for the agile project. This model generally supports empirical observations that shorter time-box iterations yield a shorter overall project completion.