Date of Award
Doctor of Philosophy (PhD)
Computer Science : Modeling and Simulation
Systems engineering., Scrum (Computer software development), Simulation methods.
Traditional systems engineering methods have been shown to not scale well to very large and complex projects. There is a need for methods that can adapt to changing technology, deliver robust solutions, and provide consistent results. Agile methods, used in the software development domain, offer one possible solution to this need. Often, new methods are developed and deployed with little examination and testing. This is partly due to the impracticality of thorough testing of engineering methods on large and complex projects. This research uses agent-based modeling and simulation to examine the suitability and advantages of Agile methods for large and complex engineering projects. The agents are modeled using Function-Behavior-Structure framework. Novel, agent-based simulations are created where the agents transition through the cognitive states of the Function-Behavior-Structure framework to develop system designs. Models and simulations are verified against data from empirical studies and are then used to examine any benefits that Agile methods have over the more traditional waterfall method. Simulations of a software design show that Scrum likely has benefits over waterfall for loosely coupled systems like software including reduction in defects, less time needed to complete the project, and greater efficiency. Scrum may have negligible benefit over waterfall for highly coupled systems with simulation metrics indicating similar performance.
Bott, Mitchell James, "An examination of the theoretical basis for agile engineering using function-behavior-structure framework and agent-based modeling" (2019). Dissertations. 179.