Navigating in Complex Process Model Collections

The increasing adoption of process-aware information systems (PAIS) has led to the emergence of large
process model collections. In the automotive and healthcare domains, for example, such collections may
comprise hundreds or thousands of process models, each consisting of numerous process elements (e.g.,
process tasks or data objects). In existing modeling environments, process models are presented to users
in a rather static manner; i.e., as image maps not allowing for any context-specific user interactions.
As process participants have different needs and thus require specific presentations of available process
information, such static approaches are usually not sufficient to assist them in their daily work. For
example, a business manager only requires an abstract overview of a process model collection, whereas a
knowledge worker (e.g., a requirements engineer) needs detailed information on specific process tasks.
In general, a more flexible navigation and visualization approach is needed, which allows process participants
to flexibly interact with process model collections in order to navigate from a standard (i.e., default)
visualization of a process model collection to a context-specific one. With the Process Navigation and
Visualization (ProNaVis) framework, this thesis provides such a flexible navigation approach for large and
complex process model collections. Specifically, ProNaVis enables the flexible navigation within process
model collections along three navigation dimensions. First, the geographic dimension allows zooming in
and out of the process models. Second, the semantic dimension may be utilized to increase or decrease
the level of detail. Third, the view dimension allows switching between different visualizations. All three
navigation dimensions have been addressed in an isolated fashion in existing navigation approaches so
far, but only ProNaVis provides an integrated support for all three dimensions.
The concepts developed in this thesis were validated using various methods. First, they were implemented
in the process navigation tool Compass, which has been used by several departments of an automotive
OEM (Original Equipment Manufacturer). Second, ProNaVis concepts were evaluated in two experiments,
investigating both navigation and visualization aspects. Third, the developed concepts were
successfully applied to process-oriented information logistics (POIL). Experimental as well as empirical
results have provided evidence that ProNaVis will enable a much more flexible navigation in process
model repositories compared to existing approaches.