Towards a conceptual framework for resilience engineering

I found this paper pretty interesting.

It’s from 2008 and provides a good primer on resilience engineering, defining types of systemic disruptions that our systems may encounter, and then some principles and heuristics that may be useful.

It’s from the infancy of Resilience Engineering as a distinct area of focus, but still holds up to today in my view.

Most interesting was the first image (that came from the end of the paper), where they provide some (tentative) resilient capacity heuristics, to guide the architecture and design in complex systems.

These heuristics include:

· functional and physical redundancy

· human backup and human-in-the-loop [* see the comments for an article discussing the problems of see-through transparency and human-out-of-the-loop automation in the case of the Boeing accidents]

· Complexity avoidance, being that systems shouldn’t be made or allowed to evolve to be arbitrarily complex (number and diversity of predictable or unpredictable interrelations, tight coupling etc.) due to inadequate design and architecture

· Graceful degradation [** Which Woods later combined into graceful extensibility]

· Drift detection and correction [** In my view, Leveson’s assumption-based indicators are a solid choice here]

In the other images, the authors discuss some common characteristics across major accidents (and almost certainly in all competitive and complex organisations), and some forms of resilience.