Toward General Principles for Resilience Engineering

Thanks to Tim Nicholas for the recommendation!


Toward General Principles for Resilience Engineering

This week we’re taking a look at some ways that ecological resilience can help to develop general principles of resilience in a paper by David J. Yu, Michael L. Schoon, Jason K. Hawes, Seungyoon Lee, Jeryang Park, P. Suresh C. Rao, Laura K. Siebeneck, and Satish V. Ukkusuri.

Some of these points may seem familiar as the authors point out, resilience engineering does address some of the points, but may not focus on them.

As you can probably guess can by its inclusion here, I see places where we software folks can benefit as well. These principles can be used to help us examine our own systems and evaluate changes to make.

The authors examine resilience from the perspective of infrastructure, both physical and virtual. They encourage a broad perspective that looks beyond a single organization or just the system that is built and raise questions like “under what set of supportive conditions or organizational abilities for learning and adaptation encouraged?”

They point out that some views of a system don’t include things outside of their immediate area, regardless of coupling or dependencies. This can influence designs in ways that may make them robust to some hazards, but especially vulnerable to others. Usually those of higher consequence and rarity. Developing more general principles may help in addressing some of these concerns.

The authors focus on “infrastructure dependent systems.” They use examples of everything from dams and levees to network systems. I like the phrase, for one because I think most of us work on systems that we would describe as infrastructure dependent. And for two because it highlights that is difficult for some systems to exist in isolation.

Themes and Principles

After examining social ecological systems research the authors develop some themes from which they then derive the principles that they suggest be included in resilience engineering research or if already present increased focus. They started with the question of “resilience of what to what” and the question of what is included in the system boundary. They focused on “sociotechnical or infrastructure dependent system in which the role of build infrastructure is clearly present” and defined the system boundary to include multilayered networks such as physical components as well as virtual and social components. With this definition could include things like dams and power plants or computerized control subway trains or even regulatory procedures and rules.

They then compared resilience engineering literature that had lists of heuristics, rules, or principles with those of social ecological literature.

Maintain diversity

This includes two types of diversity, “response diversity” and “functional diversity”. Response diversity is where multiple parts of the system may be functionally redundant, but they also respond differently to disturbances. Functional diversity is when parts of the system have different traits or do different things, the more functional diversity that exists, the more likely that parts of the system will complement each other.

While both of these are well recognized in RE research, the authors suggest that these types of diversity could also be emphasized in rules, work procedures, or regulatory infrastructure.

Manage connectivity

Connectivity is defined here as a multilayered network of built components and processes and a nested hierarchy of interacting social units that function to produce and distribute a continuous flow of essential goods and services (e.g., water, energy, mobility, etc.) for the broader society.

There are two aspects of connectivity to consider here, modularity or how compartmentalized a system is and openness, how much or little diffusion can occur across the system and it’s boundaries.

In the case of physical items, increased modularity can benefit resilience because increased compartmentalization can help prevent subsystems from transmitting disturbances to each other. Though openness is good to some degree, too much can be harmful. A system that is very open could transmit disturbances easily for example.

In the case of social parts of the system, openness can help resilience through increased transfer of knowledge or information. The authors compare too much openness in this type of system to a computer virus, where harmful or wrong information can spread easily as well.

This brings us to the notion of “fit” from social-ecological research.

“how well the structure of a collaborative social network aligns with the specifics of the environmental problem being addressed or with the structure of the biophysical system being governed”

Mismatches here with how a system is connected or modularized or governed can cause only one part of the system or issue to be addressed. The authors encourage this notion of the fit of governance and connectivity to be included in RE principles. I have been seeing increasing literature related to resilience around some of these things already.

Encourage Learning by Doing

Three “core processes” are given to help facilitate learning by doing:

  • Monitoring, where information is provided about the system and it’s environment.
  • Experimentation, small changes to a system or process where outcomes can be observed and compared.
  • Learning, where existing knowledge is updated or strategies changed based on the results of experimentation.

The previous point, around connectivity and fit can also influence learning here. Low connectivity might limit learning or limit monitoring, whereas high connectivity could provide the reverse.

The authors acknowledge that existing RE principles address learning by doing, but lament a lack of discussion on how learning should be encouraged so that it can improve adaptive capacity of systems.

Nora Jones’ work is doing much to address the issue though and has been key in helping others do the same.

Manage for Complexity

This theme really just focuses on encouraging those who govern the system or work within it to be aware of the properties of complex systems.

“The lack of appreciation among decision makers of these properties is often the reason why attempts are made to tightly regulate social-ecological systems under the idealistic assumption that these systems are tractable and predictable in the long run. This kind of command-and-control approach is attributed as the key reason why self-organizing systems gradually lose resilience”

Foster Social Capital

“The [socio-ecological] resilience literature refers to social capital in several interrelated ways—as a form of capital, as trust and leadership, as a social network, and as participation.”

This is fairly well recognized in RE, but the authors encourage a formal integration of social capital into RE principles, especially that of the role of institutional arrangements to help rapid recovery from disaster or disturbance.

Polycentric Governance

This idea where there are multiple centers of control that are interdependent tends to be represented in RE writings as “polycentric control architecture.”

The authors note that such an architecture doesn’t come without cost though, and it’s key to balance benefits of redundancy with the effort it takes to maintain. They also encouraged an increased focus on such governance structures in RE.

I was happy to see the authors point out, that this stuff isn’t easy, but developing more general principles can help:

“we hope that we have clearly communicated to building general resilience is no magic formula or blueprint panacea. Rather, we have identified principles for building resilience in the design, management, and governance of a system. Implementation is not simple, nor is success in short, however, reaching goals benefits from a clear map. Our intent has been to provide one.”

What next?

What do you think of these themes and principles? Do any of them surprise you? How are you and your organization using them today?

Takeaways

  • Examining “social-ecological resilience” can help inform how we build and think about our own systems, both as leaders and engineers.
  • Important ideas and general principles around resilience engineering can be learned from studying sociology and ecology
  • This view of resilience in eco-systems brings some things that are not yet highlighted in RE research, but may still be in the background.
  • Complex systems need to be viewed and addressed at multiple levels in order for assessments and interventions to be effective.
    • This is closely related to the notion of the “fit” of social-ecological resilience, where a system can be well or poorly aligned with the environmental problem being addressed.
  • “Risk or vulnerability trade-offs can emerge through interplays between engineering design and self-organization in the long run.”
    • Polycentric governing structure may enhance resilience or provide fertile ground for it to arise.

Don't miss out on the next issue!