Automated change impact analysis between SysML models of requirements and design

Abstract

An important activity in systems engineering is analyzing how a change in requirements will impact the design of a system. Performing this analysis manually is expensive, particularly for complex systems. In this paper, we propose an approach to automatically identify the impact of requirements changes on system design, when the requirements and design elements are expressed using models. We ground our approach on the Systems Modeling Language (SysML) due to SysML's increasing use in industrial applications.Our approach has two steps: For a given change, we first apply a static slicing algorithm to extract an estimated set of impacted model elements. Next, we rank the elements of the resulting set according to a quantitative measure designed to predict how likely it is for each element to be impacted. The measure is computed using Natural Language Processing (NLP) applied to the textual content of the elements. Engineers can then inspect the ranked list of elements and identify those that are actually impacted. We evaluate our approach on an industrial case study with 16 real-world requirements changes. Our results suggest that, using our approach, engineers need to inspect on average only 4.8% of the entire design in order to identify the actually-impacted elements. We further show that our results consistently improve when our analysis takes into account both structural and behavioral diagrams rather than only structural ones, and the natural-language content of the diagrams in addition to only their structural and behavioral content.

Predictive Models in Software Engineering: Challenges and Opportunities

“…6.2.3 Other Requirement Tasks. Apart from requirement classiication and requirement detection, several studies [22,192] built predictive models to predict and identify requirement changes and Falessi et al [58] leverage the vector space model to conduct an industrial case study for similar requirement detection.…”

A Domain-Specific Language for Simulation-Based Testing of IoT Edge-to-Cloud Solutions

“…In most of our past research with industry, we have relied on UML profiles [35] for addressing domainspecific needs. Examples include a profile for managing safetycertification documents [37] (collaboration with the maritime and energy sector), a profile for simulating tax policies [45] (collaboration with the public-service sector), a profile for traceability analysis between requirements and design [33] (collaboration with the automotive industry), and a profile for hardware-in-the-loop (HiL) testing of cyber-physical systems [42] (collaboration with the satellite industry).…”

Mining Cross-Task Artifact Dependencies from Developer Interactions

“…Jiang et al [14] apply incremental latent semantic indexing to automatically manage traceability links between code and documentation. Nejati et al [21] demonstrate the use of natural language processing to automatically identify the impact of requirements changes on system design.…”