Abstract

Project management is indispensable in many fields, including in software development. Project networks are often used to visualize the various tasks and their relations within a project. A variant of project networks, the Activity-On-Node (AON) networks, can be used to schedule the tasks of a project and, thus, determine the total duration of the project as well as which tasks can not be postponed. As a central activity in project management, the understanding and application of AON networks is an essential part of the lecture “Software Project Management” of the Research Group Software Construction (SWC) at RWTH Aachen University. In this context, the SWC Research Group uses syntactic exercise and examination tasks (ex-tasks) on AON networks to deepen and evaluate the students’ understanding. However, the creation and evaluation of such ex-tasks is time-consuming and error-prone, as there are no tools to support this process. This thesis focuses on the automatic generation of AON networks for ex-tasks. Moreover, due to the AON layout used in the lecture and ex-tasks, it is very challenging to generate such AON networks in this particular format, since known graph generation algorithms can not be used to generate graphs in the given format. We accomplish the generation and drawing of AON networks by leveraging existing approaches from graph theory and extending them for the domain of AON networks and our network structure. We develop a microservice framework to automatically generate AON networks and compute their properties. Using a microservice-driven approach, we are able to create an extensible prototype of this framework that can easily be extended in the future, to include additional features, such as automatic evaluation of ex-tasks. Using this framework, we are able to create practical ex-tasks that only need minor manual adjustments to be usable. In addition, the framework allows us to import previously created networks, calculate their properties and potentially improve their structure. Although we focus on generating relatively small AON networks, our framework is generalized enough to also handle significantly larger networks.

Resources

• Thesis report