Automatisierte Identifikation und Charakterisierung von Anforderungsabhängigkeiten – Literaturstudie zum Vergleich von Lösungsansätzen

Abstract

Many development projects of complex, interdisciplinary products fail due to an inefficient handling of requirement changes. One reason for this is that requirement dependencies and the resulting change propagation are not sufficiently recognized and considered by existing approaches. Based on a literature study and two industry workshops, approaches of automated dependency analysis of requirements for interdisciplinary development are evaluated. The result is that ontology-based approaches are particularly suitable for the automated dependency analysis of requirements due to the use of expert knowledge and the possibility of automation. On this basis, subsequent research activities can further develop the handling of requirements changes in interdisciplinary projects in a targeted manner.

Procedure Model for Structured Relational Modeling of Requirements to Support Requirements-Oriented Decision Making

“…However, the presented concepts for automation still need further development and have to be evaluated. However, it is possible to integrate existing approaches into the presented procedure (a selection of approaches can be found in (Graeßler et al, 2020)). Additionally, it must be examined, which effects occur when changes are made to the relational model and how these can be taken into account in order to maintain consistency within the models.…”

“…Especially in the development of complex systems with many requirement elements, relational modeling approaches reach their limits because of the manual modeling effort. For this reason, different approaches deal with an automated identification of relations (an extensive comparison of approaches can be found in Graeßler et al (2020)). According to Graeßler et al (2020) and Dermeval et al (2016), ontology-based approaches are a promising way to support the identification of interactions, as presented exemplary by Soomro et al (2014).…”

“…For this reason, different approaches deal with an automated identification of relations (an extensive comparison of approaches can be found in Graeßler et al (2020)). According to Graeßler et al (2020) and Dermeval et al (2016), ontology-based approaches are a promising way to support the identification of interactions, as presented exemplary by Soomro et al (2014). However, dependencies and interactions do not only occur in the area of requirements management, but also in the area of decision making, although in multi-criteria evaluation a condition is that criteria are independent of each other (Breiing and Knosala, 1997).…”

“…This information can be captured with the help of natural language processing and used for dependency analysis. Further possibilities for a supported identification of dependencies can be found in (Robinson et al, 2003) or (Graeßler et al, 2020). In the following steps of the proposed procedure, context-specific requirements for a tangible decision problem are identified by the user and the necessary relations regarding relationship, structure and condition are determined.…”

Handling of Explicit Uncertainty in Requirements Change Management

“…Available approaches to assess likelihood and impact depend on the application context. Due to interdependencies between requirements, changes might propagate (Gräßler et al, 2020). Therefore, the two risk dimensions need to be further detailed to increase accuracy of assessment and subsequent risk treatment .…”