The 10 biggest illusions in systems engineering
In the middle of the last century, when the Systems Engineering it was based on a number of assumptions. However, these assumptions were not systematically recorded, but were based on the environment at the time, particularly in the context of aerospace and defense programs. This does not correspond to today's environment.
Michael J. Pennock and John P. Wade from the Stevens Institute of Technology have investigated this issue and have drawn up a list of the 10 biggest illusions in systems engineering based on this fact. Of course, this is not the end of the story: the authors also set out a research plan to mitigate these shortcomings.
Below is the list of the 10 illusions.
The research work
So the authors have done their work Open Access articles published. This means free access to scientific literature. As I myself spend a lot of time Open Source or Open Access environment publish, I am delighted. The 8-page paper can be downloaded here.
One of the first articles at SE-Trends By the way, in December 2015, I wrote about the Stevens Institute written.
10 illusions in systems engineering
Let's move on to the actual topic, the 10 illusions in systems engineering (SE). The authors begin by discussing the history of systems engineering. They also refer to the more or less frequently discussed Identity crises in systems engineering.
Systems engineering has a clear task, a clear goal. But it is easy to lose sight of this and see it as an end in itself. This is undoubtedly the cause of some of the illusions. In addition, under certain circumstances, classic (!) systems engineering no longer works. Some practitioners do not accept the approaches that work as "real" systems engineering. Agile approaches are a good example of this.
1. illusion: Systems engineering is absolutely
Traditional SE assumes that there is a "right" or optimal solution. The task is really to optimize and apply best practices, rather than being a subjective and context-dependent process. This is one of the key assumptions that often separates traditional systems engineering from systems thinking, which takes into account the soft and often inconclusive nature of systems.
2. illusion: Systems engineering is clear
Systems engineering is based on the illusion that it is possible to create unambiguous Requirements in human language. Of course, experience teaches us that people can interpret a requirement in very different ways. To make matters worse, we expect these requirements to be decomposable into further, subordinate requirements.
3. illusion: Systems engineering is sequential
Traditional SE requires a fundamentally sequential approach to development. This does not mean that there are no feedback loops, but generally the system to be developed is such that its development can be decomposed in time and carried out sequentially. This makes it difficult to develop systems that need to be adaptable.
4. illusion: Actors in systems engineering act rationally
We often assume that decision making in engineering is similar to that of business actors: Engineers and managers are rational decision makers who have access to complete and perfect information. In real life, systems engineers are under deadline and political pressure as well as resource constraints. As a result, they often do not have the information they need to make informed or "optimal" decisions. In the face of such ambiguity, the seemingly best option may be to do things the way they have always been done. However, in the case of complex systems or systems of systems, this can be exactly the wrong approach.
5. illusion: Systems engineering leads to minimalist, loosely coupled systems
The key to systems engineering is the ability to divide and rule. But for this to work, the system in question must be decomposable into reasonably loosely coupled subsystems. With increasing Complexity minimalism becomes problematic, as the subsystems become increasingly artificial and potentially counterproductive from a technical point of view, as critical interactions cannot be adequately taken into account or recognized.
6. illusion: Systems engineering is controlled centrally
Control over the development organization(s) is centralized and absolute. Traditionally, this was regulated by contracts and similar mechanisms. With systems of systems, however, there are often no explicit contracts. As a result, it is more difficult for the systems engineer to maintain technical control. Subsystems can change unexpectedly, which can lead to problems and failures.
7. illusion: Systems are unchangeable
Traditional SE assumes that the system requirements and the system context essentially do not change. In the real world, of course, competitors behave unexpectedly, new technologies are introduced, suppliers go out of business, etc. As social and technological change accelerates over time, static solutions may become obsolete before they are deployed.
8. illusion: Systems engineering is mechanistic
Traditional systems engineering is blind to the human element, including culture and history. There is an implicit assumption that these factors do not significantly affect outcomes. However, norms and values can affect what potential solutions are considered and how a systems engineering program is executed. The potential adverse effects that an unsuitable organizational structure can have on system developmentare generally known.
9. illusion: Systems engineering is deterministic
Traditional systems engineering assumes that system behavior is deterministic. The input "A" in combination with the state "B" always results in the output "C" (or at least captures the probability of C with a precisely defined probability distribution). This view becomes problematic for adaptive systems where a certain amount of fault tolerance is required in exchange for the benefits of adaptability, Keyword AI.
10. illusion: Systems engineering is context-free
Best practices are universal and do not vary according to context. In other words: If a certain systems engineering approach has proven itself in the aerospace industry, it must also work well in the IT sector. Not at all! This view neglects the importance of tailoring. Different aspects of the systems engineering process may be more or less important for different types of systems and industries.
How to deal with the illusions?
Pennock and John Wade have drawn up a research plan based on these illusions (in the paper, Chapter 5, Research Agenda). That may be all well and good in an academic environment, but how does it help us in practice?
It's not that difficult: the most important thing is to recognize these illusions in the first place. Because if we are aware of them and have created an awareness of them in the project, then we are already a good deal closer to our goal.
Photo by 愚木混株 cdd20 on Unsplash
