Due to a continuously increasing process complexity in any field of business nowadays it is necessary to create methods to display, evaluate and improve processes. The direct impact of entities (e.g. tasks) onto their immediate neighbours is a good approximation to identify entities relevant to the process. Therefore, metrics to evaluate the degree of such a node in a graph have gained importance. Kreimeyer et al. developed a way to model processes with a matrix-based notation including logical operators as AND, OR and XOR by using a modified Multiple Domain Matrix (MDM). The approach integrates different complexity measures to assess process structures using the Active and Passive Sum metric (also called 'Activity' and 'Passivity'), McCabe's Cyclomatic Number, and Cardoso's Control-Flow Complexity. However, so far, the use of these metrics has not been demonstrated in depth. This paper closes this gap by presenting a case study that makes use of the Active and Passive Sum extended to include Boolean operators. All achievements shown in this work are utilized on a case study of a big German automobile manufacturer, including 160 tasks, 134 Business Objects and 54 OR-connectors. The slides to this abstract present this case study using the principles shown in the short generic example. Besides the general functionality, also the disparity between AND- and OR-connections towards their structural impact was shown. High Active / Passive Sums were especially found for AND-connections, although OR-connections often visually look more 'dramatically' but impact the structure to a lesser extent. However, this fact should not be underestimated as, of course, the possibility of an impact is still given; therefore, in a second step, the interpretation of the results obtained through the extension of dependency modelling also needs to be adapted. However, the adapted metric is based on the assumption that all eventuating cases happen with equal probabilities. This might not necessarily be the case; yet, from a structural point of view it is still closer to the behaviour of a system compared to replacing all decision points with AND operators. Therefore the presented approach allows a more substantial evaluation based on the actual dependencies in the network. All calculations can be automated, thus facilitating the analysis for large structures. Combined AND-OR-connections and iterations are solved as well as ill-designed processes (i.e. process models that do not adhere to modelling conventions, e.g. tasks being connected to themselves or missing business objects in between tasks).
Extending the Active Sum / Passive Sum measure to include Boolean operators: A case study
Erweiterung der Aktiven Summe / Passiven Summe Messung, um Boolean Operatoren einzubeziehen: Eine Fallstudie
2009
14 Seiten, 3 Bilder, 1 Tabelle, 6 Quellen
(davon 10 S. Folien)
Conference paper
Storage medium
English