Multi-Dimensional Process Analysis
Processes are complex phenomena that emerge from the interplay of human actors, materials, data, and machines. Process science is the discipline of studying and improving processes by developing effective methods and techniques to do so. The field has developed mature theories, models, and solutions for studying and improving process executions from the control-ow perspective. Also, the limitations of control-ow focused thinking are well-known in the field, leading to active involvement of various disciplines to study phenomena “beyond” control-ow. However, it remains challenging to relate models and techniques of various fields to the dominant control-ow oriented paradigm. This tutorial introduces several recently developed simple models that naturally describe phenomena beyond control-ow, but are inherently compatible with control-flow oriented thinking. We discuss the Performance Spectrum to study performance patterns and their propagation over time, Event Knowledge Graphs to study networks of behavior over data objects and actors, and Proclets as formal model that allows reasoning across control-ow, data object, queue and actor behavior. For each model, we discuss which phenomena can be studied, which insights can be gained, which tools are available, and to which other fields they relate.
Dirk Fahland is an Associate Professor in Process Analytics on Multi-Dimensional Event Data at Eindhoven University of Technology (TU/e). His research area is the analysis and improvement of complex, distributed systems through event data, process mining, and explainable models. Dirk has contributed to research in process management and mining since 2008 in over 80 journal, conference, and workshop publications with foundational results in process modeling, discovery, analysis, and repair. His current research specifically studies cause-effect relations and emergent behavior in networks and dynamic systems as a whole. Insights gained in numerous industrial projects led to the idea of encoding behavioral information in event knowledge graphs, a cornerstone of a new generation of “Augmented BPM Systems”.