TL;DR: I propose that Systems Engineering be defined as ‘The study of applying systems science and systems heuristics to the conceptualization, design and development, integration, verification and validation, operation and sustainment, retirement and renewal of a set of actants and relationships between those actants.’ This article is how I go there.

Here’s an idea, let’s take Systems Engineering at its face value: the engineering of systems. The output of such an activity is of course, at least, one or more engineered systems.

Let’s have a look under the hood at what is ‘engineering of systems’.

Engineering: ‘the study of using scientific principles to design and build machines, structures, and other things, including bridges, roads, vehicles, and buildings’ – Oxford Dictionary.

System: ‘a set of connected things or devices that operate together’ – Oxford Dictionary.

If we combine the the terms then Engineering of Systems is ‘the study of using scientific principles to design and build a set of connected things or devices that operate together’. However this definition has several limitations: process, constraint, and subject.

Process – Application

The definition takes engineering to be ‘The study of using scientific principles …’. What’s valuable to recognize is that the definition is stating that engineering is the study of using xyz, not directly using xyz. This is pertinent because engineering is rightly rooted in the Plan and Check part of the Plan-Do-Check-Act cycle as opposed to the Do and Act part. Of course, verifying and validating engineered artifacts is critical to successful engineering but engineering doesn’t start at that point; it starts a lot earlier. Moreover, I’m not keen on ‘using’ in this context, and prefer ‘applying’ instead. ‘Using’ is the more generic term, but ‘applying’ is the more active one, and I’d argue engineering is the application of science, not merely the use of it (using science, after all, sounds more like what scientists themselves do).

I also believe these more specific terms should replace the generic ‘scientific principles’ when engineering is applied to systems, precisely because systems science is a relatively young field, and systems heuristics can help fill that gap. My motivations to do so are two-fold: firstly, to recognize that science applied to systems is still in its infancy and as such heuristics can cover for this gap in rigorous scientific proofs; and secondly to draw attention to the existing, growing and increasingly important field of systems science.

I therefore propose to transform ‘The study of using scientific principles …’ to ‘The study of applying systems science and systems heuristics …’.

Constraint – Phases

The definition is further limited to the design and development phase. It neglects the phases that a system goes through which are Concept Definition; Design and Development; Integration, Verification and Validation; Operation and Sustainment, Retirement and Renewal (see SEBOK). If we take as a premise that the engineering of a system should consider all phases of the engineered system’s lifecycle then it is reasonable to expect that the definition should reflect this expectation.

I then transform ‘… to design and build …’ to ‘… to the conceptualization, design and development, integration, verification and validation, operation and sustainment, retirement and renewal …’.

Subject – System

The previous offered definition has a few terms that are redundant with each other, namely ‘things’ and ‘devices’. As far as generalizations go, ‘things’ is as generic as you can get, it could refer to the Solar System or a rock sitting on your desk, or even ‘devices’. As ‘devices’ is a subset of ‘things’ the Oxford Dictionary definition is cumbersome and would be similar to saying ‘a highway is for the transport of vehicles and cars’. I am also of the strong opinion that systems engineering should consider the human actor as an integral part of the system. As people are clearly not devices and referring to people as things is technically correct but rude I offer to dismiss the use of ‘things’ or ‘devices’ in favor for the term ‘actant‘.

Actant: ‘an entity that can act to influence a network, anything that can modify other entities through action’ – The Cynefin Company.

What I like about actant is that it speaks to something that is not only connected to others in a network but also that it has an impact on others through action, where that impact could be a change in characteristics (e.g. mass) or change in behavior of that connected thing. It is also vague enough to refer to actors, devices or even concepts. However, swapping out ‘things and devices‘ with actants in the definition wouldn’t be enough because it implies that the system is again limited to the operational phase of that system (‘… that operate together’). Whilst keeping the concept of set in the definition I would extend that to the set of actants and relationships between actants. Doubling down on the concept of set it also opens the door to a mathematical definition of a system which I propose as follows:

S = (A​, R)

where A = {A₁​, A_2​, …, A_n​}

and R ⊆ A × A

such that R₁₂​ : A_1​ → A₂​

where S is for System, A is for Actant and R is for Relationship. Note that I’m not assigning a purpose or teleology to the revisited definition of system as I don’t subscribe to the notion that a system has such a characteristic outside of the idea that the system is a set of actants and relationships between those actants.

I therefore propose to transform ‘… a set of connected things or devices that operate together’ to ‘… a set of actants and relationships between those actants.’

Conclusion

I then draw everything together.