2. State of the art and related research

DDMBSE is an evolution of traditional MBSE that integrates DA methods into the modeling process. The integration of DA into MBSE is viewed as one of the core directions for the successful introduction of MBSE (INCOSE, 2021) and following the diverse DA application areas such as supply chain management (Reference Schoenherr and PeroSchoenherr and Speier-Pero, 2015) or fraud detection (Reference Vassakis, Petrakis and KopanakisVessakis et al. 2018). DDMBSE enhances system models by incorporating real-world data, enabling continuous updates and feedback loops. Initial research in defining DDMBSE has been conducted by Tissen et al. through a literature review (Reference Tissen, Wiederkehr, Bernijazov, Koldewey and DumitrescuTissen et al. 2023) and an interview study (Tissen et al. Reference Tissen, Wiederkehr, Bernijazov, Koldewey and Dumitrescu2024b). DDMBSE relies on standardized interfaces to connect diverse data sources, such as sensors, databases, and simulation tools, with system models, transforming static representations into dynamic, data-driven frameworks. By fostering advanced traceability and automation, it enables more efficient handling of system complexity and accelerates development processes (Reference Tissen, Wiederkehr, Bernijazov, Koldewey and DumitrescuTissen et al. 2023). Furthermore, DDMBSE supports decision-making through data-driven insights, improves system model quality through real-time updates, and reduces errors by enabling early validation and verification of requirements. However, challenges remain, including data preparation, ensuring data quality, and fostering user acceptance. Successful implementation requires targeted training, organizational alignment, and the development of standardized tools and methods to support interdisciplinary collaboration and scalability across industries. Additionally, the establishment of a clear data strategy and the integration of robust data pipelines are critical to achieving the full potential of DDMBSE (Tissen et al. Reference Tissen, Wiederkehr, Bernijazov, Koldewey and Dumitrescu2024b). Therefore, a maturity model has been developed (Tissen et al. Reference Tissen, Bernijazov, Koldewey and Dumitrescu2024c). Existing literature in the fields of MBSE and DA has begun to explore the categorization and description of use cases. Studies on MBSE often focus on domain-specific applications, such as aerospace or automotive systems, providing examples of how modeling techniques are applied. Xames and Topcu have investigated and systematized the use of MBSE in the healthcare sector (Reference Xames, Md and TopcuXames and Topcu, 2024), while Wenyue et al. have examined the use of MBSE in aerospace (Reference Wenyue, Junjie, Yinxuan and ZhiangWenyue et al. 2022). In contrast, a variety of use case collections and methods for describing use cases have been published in the area of DA. Kühn et al. developed a canvas for the technical specification of DA use cases in an industrial context, without providing an overview about existing use cases (Reference Kühn, Joppen, Reinhart, Röltgen, von Enzberg and DumitrescuKühn et al. 2018). Meyer et al. proposed a framework for defining and specifying DA use cases in the context of strategic product planning, but did not mention the product development process or MBSE in particular (Reference Meyer, Panzner, Koldewey and DumitrescuMeyer et al. 2021). Kaiser et al. (Reference Kaiser, Schräder, Bernijazov, Foullois, Dumitrescu, Koch, Wilke, Dreiseitel, Kaffenberger and Gesellschaft2022) describe an approach in structuring AI-assistants in MBSE along model-oriented and process-oriented categories, but not directly to single use cases since AI-assistants can fulfill multiple use cases and are software-based. In the intersection of DDMBSE, no attempt to single use case categorization and description is known to the authors. Led by the research question “How can DDMBSE use cases be systematically categorized, organized, and described?”, this paper addresses this gap by proposing a categorizing framework that provides a comprehensive approach to organizing use cases of DDMBSE.

3. Research design

Guided by the research question posed in the preceding sections, the overarching methodology of design research (DRM), as adapted from Manson (Reference Manson2006), was employed to develop a categorizing framework for DDMBSE, which included the creation of a use case specification template (see Fig. 1). The overall research design consists of five process steps (awareness, suggestion, development, evaluation, conclusion):

1. 1) Awareness: Section 1 and 2 begin with an examination of the factors that have led to the proposal of new methodologies and best practices in the context of DDMBSE. They represent process step 1 (Awareness) in the design research methodology.

2. 2) Suggestion: The second process step (Suggestion) is the proposal of a structured use case collection in the form of a clustering framework. Previous studies by Tissen et al. identified use cases for DDMBSE based on their results from a systematic literature review (SLR) (Reference Tissen, Wiederkehr, Bernijazov, Koldewey and DumitrescuTissen et al. 2023) and an interview study with experts from research and industry (Tissen et al. Reference Tissen, Wiederkehr, Bernijazov, Koldewey and Dumitrescu2024b). Given the accelerated rate of progress in data-driven research fields, an updated SLR is conducted to provide a comprehensive overview of the current state of the art in addition to the previous studies. The database SCOPUS was selected because of its extensive repository of peer-reviewed research proceedings, journals, conference papers, and articles. The search was also compared with results based on Web of Science. However, due to the overlap and the overall greater number of results, SCOPUS was chosen for the following stages of the procedure. The search string, defined as (“systems engineering” OR “model-based systems engineering” OR “MBSE” OR “systems thinking”) AND (“data” OR “data analytics” OR “data-driven”), yielded a total of 20,066 papers. We excluded terms like artificial intelligence and machine learning since the SLR after Tissen et al. Reference Tissen, Wiederkehr, Bernijazov, Koldewey and Dumitrescu2023 did not include them. Following the application of the inclusion and exclusion filter (Years: >2012 < 2025; Subject area: computer science, engineering, multidisciplinary; Document type: conference paper, article, conference review; Source type: conference proceedings, journals; Source title (exemplary: Procedia Computer Science, IEEE Access, Procedia CIRP, …); Language: English)) and analyzing Title, Abstract and Keywords (supported by ChatGPT and then proofread), 60 relevant paper were identified. The criteria for selecting suitable source types was oriented by whether the source type is thematically relevant for MBSE or DA. The filtering of on the level of Title, Abstract and Keywords was guided by the question, whether the main results of the paper give a hint into a possible DDMBSE use case or not. A use case here represents an aspect, artefact, method etc. based on MBSE, which in some beneficial form is involved or extended in its kind by data-driven techniques, methods, tools or aspects etc. Following a comprehensive analysis of the full paper and the incorporation of additional sources from previous knowledge, a total of 35 papers were identified as forming the basis for the extraction. Each paper was then analyzed led by the search question: What are use cases of MBSE where data analytics methods are applied?. With these criteria in mind, 47 relevant use cases could be identified, resulting in an overall total of 109 use cases.

3. 3) Development: During the third process step (Development), the use cases and source materials were analyzed and reviewed, together with the preceding structures. As the use cases from Tissen et al. (Reference Tissen, Wiederkehr, Bernijazov, Koldewey and Dumitrescu2023/2024) have already been clustered, the high-level clusters were used as one axis. They are described as high-level categories, which resulted in logical derivates. Since a use case follows some possibility or need that it fulfills, searching for a use case implies that the user has some form of interest in a use case. Accordingly, the second axis was defined with the general concerns of systems engineers and data analysts - the major user groups in context of DDMBSE - towards DDMBSE (Tissen et al. Reference Tissen, Wiederkehr, Bernijazov, Koldewey and Dumitrescu2025). It comprises 22 aspects distributed across the four high-level groups: optimization, support, detection, and understanding. Given the evident interconnection between the application of a use case and the associated concerns, this structure appeared to be a logical choice. Following a period of testing and refinement, the 109 use cases were sorted by two researchers and the framework was constructed. Furthermore, a template was designed to describe each use case in the context of DDMBSE, providing all the necessary information. This includes a detailed description of the use case, its objectives, context, constraints, involved stakeholders, addressed life cycle phases and data information.

4. 4) Evaluation: Subsequently, an exemplary use case was employed to illustrate the sorting process within the framework, with the aid of the proposed use case template (Step 4: Evaluation). The final process step represents the conclusion of the DRM as outlined by Manson (Reference Manson2006). It involves a discussion of the results obtained from the designed artefacts (the framework and template) and an analysis of their alignment with the initial process step (Awareness) and the defined requirements. Section 4 describes the results of process step 3 in detail, followed by section 5 with the demonstration of the use case template on one use case based on a completed research project in context of DDMBSE.

5. 5) Conclusion: Process step 5 then discusses the findings and concludes this paper.

Figure 1. Overall research design (adapted after Manson (Reference Manson2006)) and detailed steps (UC: use case)

5. Demonstration

This section is concerned with the demonstration of the designed artefacts, which have been created in accordance with the structuring framework for DDMBSE use cases and the description template (Step 4 - Evaluation in the DRM). To illustrate, a completed research project called iQSketch is employed as a case in point (see Fig. 4). iQSketch represents use case 51 (digitalisation of workshops), within the category of tool usability and assistance, and addresses concerns pertaining to economic optimisation, the expansion of modelling methods through data and AI, data handling and data linking and consistency. The sorting is based on the objective of the use case, which is to automate the process of digitizing informal images of workshop results and transferring them into a formal digital representation. The main problem behind the use case is that current practice in MBSE workshops is to use sticky notes and brown paper to create initial system models and diagrams. This process is time-consuming and resource-intensive, particularly when the results must be transferred manually into a modelling software tool. In this context, image recognition and object detection models must be employed and trained to analyse a variety of diagram types, thereby necessitating a higher level of DDMBSE maturity, whereby MBSE is utilised in conjunction with data analytics departments. The project involves a number of different stakeholders, including systems engineers, product managers, data engineers, software engineers and domain specialists. The data employed within the use case is derived from images captured at various points in time from workshop results or comparable tools and subsequently transformed into formal data formats, such as JSON. Given that the use case is employed at an early stage, the addressed lifecycle phases are situated within the concept and development phase. The template allows the user to carefully describe the essential points of a future or already implemented DDMBSE use case, highlighting all necessary information to describe, understand and implement it.