Complex Behavior Emerges from Simple Rules

At TEDxZurich 2013 Nicolas Perony gave an excellent talk about complexity that has relevance in the collaboration technology field.  He has some excellent examples of how complex systems differ from complicated, machine oriented system, as well as the risks of thinking of complex systems as if they were machines.

The more complicated you make a machine, the more likely it will fail due to something absolutely unexpected.

Puppies: You have to see the video of this excellent example of emergent behavior.  It shows six hungry puppies interacting as they normally do: randomly.  However, when milk is given to them in a single bowl, an excellent example of emergent behavior is produced.  The puppies push on each other to get to the bowl, and they start to circle the bowl in a pinwheel pattern.  Clearly, nobody instructed the puppies to do this, nor is there any evidence that they know they are doing this, yet it simply emerges from the puppies enthusiasm to get to the milk.

Complex ≠ Complicated: His description of complicated versus complexity is very good and very clear: “Something complicated comprises many small parts, all different, and each of them has its own precise role in the machinery.  On the opposite, a complex system is made of many many similar parts, and it is their interaction that produces the globally coherent behavior.”   This is well depicted in this graphic:

ComplexComplicated

He goes on to say that each of the individual elements of a complex system operates by (relatively) simple rules, but — and this is the crucial part — the emergent behavior of the system as a whole can not be predicted by the individual rules alone.

Robot Failure: He given an example of a robot he built to follow a line on the floor.  IT was a very complex machine which failed when demonstrated in a different room than it was developed in.  A very small unexpected change in the lighting cause the entire operation to fail to follow the line.  The more complicated, the more liable to fail.  How is it then that biological system achieve what they do so well.

Bats: He studied how a bat colony exhibits very sophisticated behavior.  How do they do this.  It turns out that the rules regarding the behavior of a single bat might be quite simple.  They made a computer model with these simple rules that explained the complex behavior of the colony.  (The rules are simpler, but remember the rules don’t represent the behavior of the colony at all.)

Meerkats: He studied the behavior of meerkats and found that meerkats tend to follow a particular dominant meerkat, until they reach a barrier, such as a road.  At that point, 80% of the time, a non-dominat meerkat crosses the road first.  How is this accomplished?  How does the dominant meerkat tell the subordinant to go first?  It doesn’t.  Again, they found some simple rule in the individuals, that produced the emergent behaviors.

Animals show extraordinary social complexity, and this allows them to adapt and respond to changes in their environment.
Simplicity leads to complexity which leads to resilience.

Take Away

Like an ecosystem or social behavior of animals, human organizations also display very complex social behaviors.  This is the goal of the Business Process Management discipline to support those processes.

The standard, engineering approach is to analyze the global emergent processes, to codify them, and to automate them.  However this treating the organization as a machine with specific coded responses for specific stimuli is fragile:  a small change in the environment can break it.

Perhaps what we think of a business processes, are really epiphenomena that result from internal rules on individuals that are more or less unrelated to the resulting emergent behavior.  If that is true, we should focus on thore individual “micro-rules”.  This would yield a system that is far more resilient to external changes.

But how to do that?  Nobody — as far as I know — is working on deriving such repeatable local scale rules that would, through interactions with large numbers of individuals with similar rules, produce business processes as an emergent behavior.  Is this the future for BPM?

Advertisements
This entry was posted in Adaptive, Adaptive Case Management, complexity and tagged , , , , . Bookmark the permalink.

2 Responses to Complex Behavior Emerges from Simple Rules

  1. Hi Keith, great summary and references! While I am glad about a TED speech on the subject, it is really interesting that it took so long for the mainstream to grasp the difference of complicated versus complex.This subject was the reason we ran into each other. I remember that your response to my following post was: ‘I think you are on to something here.’

    http://isismjpucher.wordpress.com/2009/05/24/bpmn-2-0-a-step-sideways/

    I have been following the research of complex adaptive systems coming at the Santa Fe Institure since the late 80s. So we are in fact nearly 20 years later and only slowly the understanding creeps in. That post was written in the middle of 2009, several months before we named the different approach to supporting process collaboration ACM Adaptive Case Management.

    I am sorry to say that the experts in BPM, management consulting and business strategy are still stuck in scientific management concepts, but maybe it has to come top down and not bottom up. I have the most amazing discussions in process management groups where the focus is still to implement as many best practice and standard processes as possible. Phew … quite tiring.

    Let me finally comment on the idea of ‘social local rules’ that in the large support business processes in a resilient manner. There is no need for BPM to capture those rules because they are neither repeatable nor can they be automated. These rules follow the basics of human interaction and in the large it is the diversity of people in a business and not the commonality that produces the emergence and resilience effects. All process management has to do is to support human interaction and help management to empower people with authority, goals and means. ACM the way I see (and try to promote) it provides the IT support structure for just that.

    Thanks again for hanging in there. We still have some way to go!

    • kswenson says:

      I remember that conversation well. We have come a long ways. HBR had that issue in Sept 2011 on complexity. Last fall, the Global Peter Drucker Forum had their yearly conference with the theme this year “Managing complexity” with a couple of members of the Santa Fe Institute participating. Complexity is starting to get into the mainstream zeitgeist, however there is still a long ways to go.

      The “micro-rules” indeed follow the basics of human interactions, but we never actually look directly at those micro rules. We tend to focus exclusively on the macro patterns, and somehow forcing those to exist. It would be like forcing snowflakes to be hexagonal by using molds or special cutting equipment, while ignoring the way that atoms pack themselves into a crystal. If we were to focus only on human interactions, and if we had a way to regulate those micro rules, and if we had simulation to try out different micro-rules to see which worked better, we might be able to make business processes that work without having any idea in advance precisely how a particular instance is handled. We know this can be improved by Darwinian style selection, but most business people would think that would be very inefficient. My question: would it be less efficient than what we do with the brute force macro-scale way?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s