MBSE in research
What topics in the field of Model Based Systems Engineering (MBSE) are currently being researched? An international group of researchers, led by Junda Ma from the Beijing Institute of Technology, investigated this question. Using a literature search, they answered a series of questions about MBSE in research, with some surprising results, particularly with regard to tool chains.
The paper Systematic Literature Review of MBSE Tool-Chains was published in Applied Sciences in 2022. In addition to Junda Ma, Guoxin Wang and Yan Yan, also from the Beijing Institute of Technology, as well as Jinzhi Lu and Dimitris Kiritsis (École Polytechnique Fédérale de Lausanne) and Hans Vangheluwe (University of Antwerp) also took part.
The analysis is based on a total of 260 publications on the topic of MBSE.
Questions
The researchers set themselves the goal of answering the following four questions:
- How much research on MBSE tool chains is actually being carried out? This question was intended to capture the scope of research topics related to MBSE tool chains.
- What are the most frequently investigated research areas in connection with MBSE tool chains? And in which areas of application were they used? This research question was used to analyze how MBSE tool chains support the work of researchers in systems engineering and in which areas they are used.
- What are the most commonly used research methods? With this question, the researchers wanted to understand which techniques can be used to build MBSE tool chains. Four aspects were considered, based on the SPIT Framework for the development of tool chains. The SPIT framework was presented at the INCOSE Symposium in 2017.
- The last question dealt with the contribution of research in connection with MBSE toolchains.
From 0 to 30 in 20 years
Zero is perhaps an exaggeration, but until around 2007 there was very little research into MBSE tool chains. This may be somewhat misleading, as there was already a certain "modeling boom" at the time. But back then, the industry was focused on software modeling in general and UML in particular.
Areas of application
In the second question, the researchers examined the stakeholders of MBSE and their roles, as well as the development phases in which MBSE is used.
The second question in particular certainly does not reflect the actual use of MBSE, as research papers were specifically examined. Accordingly, MBSE was also used in just under 60% of the papers for research projects.
More interesting for us is the analysis of the stakeholders. In a good half of the papers, this is the systems engineer. What is surprising, however, is that the domain engineer is the stakeholder in a third of the cases. In next place comes the Project Manager with only 5%.
Alongside the systems engineer, the domain engineer is one of MBSE's biggest stakeholders.
Systematic Literature Review of MBSE Tool-Chains
The researchers also investigated the domains for the use of MBSE. There were hardly any surprises here: Aerospace and automotive dominate, with the full spectrum of industries represented. Almost 50 of the papers examined did not refer to any particular domain.
Functional safety technologies dominate
Question 3 deals with the specific techniques that can be supported with MBSE tool chains, such as reliability, scalability and the like. Not surprisingly, functional safety techniques such as safety, reliability, resilience and quality dominate, together representing just under 50%.
The intended use of MBSE is probably even more interesting for managers and decision-makers. The abstract understanding of problems dominates (just under 50%). Domain-specific modeling was in second place, which is in line with stakeholder analysis.
Modeling languages: SysML and UML
SysML and UML together represent 75% of the modeling languages used. Also worth mentioning are the domain-specific modeling languages (DSML), which were used in as many as 13% of the cases.
The tools used are exciting, but should also be used with caution. This certainly does not reflect industrial use, but is specific to research. Nevertheless, the students of today are the industrial users of tomorrow.
Many of these tools are so flexible that they can be used for different activities. The cover image of this article shows how the tools are distributed across the five application areas of requirements management, system modeling, analysis, verification and project management.
Quo vadis, MBSE?
The last question examined the focus of MBSE research. For practitioners at least, it is hardly surprising that these are frameworks or methods with a whopping 60%. In next place are tool chains with just under 12%. This also reflects my practical experience. In these areas (methods, frameworks), the level of suffering in the industry is chronically very high, and cooperation in the research project promises to remedy this.
I have become disillusioned with tool development in research: The tools are rarely suitable for practical use because the doctoral students work on topics that they enjoy (or that are important for the dissertation) and not on aspects that are boring but essential for practical use.
Conclusion: I find literature analyses like this very helpful, especially for us practitioners, in order to maintain an overview. Many results confirm my impression of the market, for example the dominance of SysML. On the other hand, the biggest surprise for me was the dominance of domain-specific aspects, which we saw in the stakeholders, problems and languages.
If you are interested, I can recommend that you take another look at the paper yourself.
Image source: Systematic Literature Review of MBSE Tool-Chains (CC-BY)
