Complexity and Understandability Comparison between Unstructured and Structured Business Process Models

Raul Mäesalu
The main goal of my work will be to determine if an unstructured business process model (BP model) is more understandable after being converted into an equivalent structured variant. In order to reach that conclusion I will study a dataset of models from real-life composed by IBM. I will use an open-source program called Bpstruct to generate structured equivalents for all models in the two datasets that qualify as structurable. Then I will analyze the models (unstructured models and their equivalent structured versions) using complexity metrics introduced in the academic literature on BP models. Using these measurements I will perform statistical analysis in order to compare the properties of unstructured models and of their equivalent structured versions. From this analysis, I will also select some sample models that will illustrate the complexity tradeoffs between unstructured and structured models. The extracted sample models will be used to conduct a survey among students in the University of Tartu who attend the Business Process Management course to empirically evaluate if the structured variants of models are simpler to understand. The results of the survey will be combined with the measurements of the models to reach conclusions about whether structuring an unstructured BP model makes it simpler and easier to understand.
Graduation Thesis language
Graduation Thesis type
Master - Information Technology
Marlon Dumas, Luciano García-Bañuelos
Defence year