Herd Structures and Complex Adaptive Systems

Lessons from ants and The Walking Dead

Published: Tuesday, February 5, 2019 - 13:03

The TV show The Walking Dead, about survival in a post-apocalyptic zombie world, is one of the top-rated currently. I’ve written previously about the show, but today I want to briefly look at the complex adaptive systems (CAS) in the show’s plot structure. A CAS is an open, nonlinear system with heterogeneous and autonomous agents that have the ability to adapt to their environment through interactions between themselves and the environment.

The simplest example of a CAS is an ant colony. Ants are simple creatures with no leader telling each ant what to do. Each ant’s behavior is constrained by a set of behavioral rules that determine how it will interact with others and its environment. Each ant works with local information and interacts with other ants and the environment based on this information. Their tasks include patrolling, foraging, maintaining the nest, and performing midden work. The local information available to each ant comes from the pheromone scent from another ant. As a whole, their interactions result in a collective intelligence that sustains their colony, a complex and intelligent system.

In response to perturbations in their environment, ants are able to switch to specific tasks to maintain their system. Each ant decides its task based on the local information, in the form of perturbation to their environment, as well as the ant’s rate of interaction with other ants performing specific tasks. In the presence of need, ants can rise through the ranks, eventually becoming foragers, a position they maintain for the rest of their lives. The ant colony supports a large amount of “reserve ants” that don’t perform any function. This reserve allows for specific task allocation as needed based on perturbations to their environment.

To further illustrate the “self-organizing” or pattern-forming behavior of ants, let’s take their foraging activity as an example. Ants will set out from the colony in a random fashion, looking for food. Once an ant finds food, it will bring it back to the nest, leaving a pheromone trail on its way back. Other ants engaged in foraging will follow the pheromone trail, bring back food, and also leave their pheromone scents on the path. The scent evaporates after a short amount of time.

Ants that followed the shortest path will leave a pheromone trail that stays “fresh,” while the trails on a longer path won’t remain as fresh because the pheromone has more time to evaporate. This means that the path with the strongest pheromone trail is the shortest path to the food. The shortest path is a result of positive feedback loops from more and more ants leaving pheromone at a faster rate. Here, the local information available to each ant is the rate of pheromone release from the other ants. The faster the rate, the stronger the trail. This generally corresponds to the shortest trail to the food source. Once the food source is consumed, another food source is identified, and a new short path is established.

As an aside, several transportation companies have adapted this “ant colony optimization algorithm” to find the shortest delivery routes.

In The Walking Dead, a similar collective behavior is shown by the zombies. They exhibit a herding behavior where a large number them will move together in search of “food.” Similar to ants, they are devoid of any intelligence, and there is no one in charge. The zombies are attracted to sound, movement, and, since they don’t attack each other, possibly an absence of “zombie smell.” In fact, in the show several characters were able to survive a zombie attack by lathering themselves in the “zombie goo.”

The possible explanation for the formation of herd structures is the hardwired attribute that we all have—copying others. We tend to follow what others are doing when we are not sure what is happening. We go with the flow. We could develop a model where a few zombies are attracted by a stimulus, and they walk toward it. Other zombies simply follow them, and soon a large crowd forms due to the reinforced loops with more and more followers. This is similar to the positive reinforcing feedback of a pheromone trail laid down by ants.

The show recently introduced an antagonist group called “Whisperers.” They worship the dead, adorn themselves with zombie skins, and walk among the zombies. They learned to control the herd and make it go where they want. Possibly, they are able to guide the zombies by first forming a small crowd themselves and then getting zombies to join them. Because they have the “zombie smell” on them, the zombies don’t attack them. The Whisperers, themselves a CAS, have adapted by joining the zombies.

How does understanding complex adaptive systems (CAS) help us?

We’re not ants and certainly not zombies. But there are several lessons we can learn from understanding CAS. The underlying principle is that we live in a complex world that we can understand only in the context of our local interactions with neighbors and the environment. We all belong to a CAS, whether at work or in our communities. Every project we are involved in is new in some way; it could be the nature of the project itself, the team members, the deadlines, or the client. Every part of a project can introduce a variation that we didn’t know about.

Here are some lessons from CAS that can help in project management.

1. Observe and understand patterns.
CAS present patterns due to the agents’ interactions. You must observe and understand the different patterns around you. How do others interact with each other? Can you identify new patterns forming in the presence of new information—or perturbations in your environment? Improve your observation skills to understand how patterns form around you. Look and see who the “influencers” are in your team.

2. Understand the positive and negative feedback loops.
Observe and understand the positive and negative feedback loops that exist around you. A pattern forms based on these loops. An awareness of the positive and negative loops will help us nurture the required loops.

3. Be humble.
Complexity is all around us, and this means that we often lack understanding. We can’t foresee or predict how things will turn out every time. Complex systems are dispositional, to quote Dave Snowden. They may exhibit tendencies, but we can’t completely understand how things work in a complex system. Edicts and rules don’t always work, and they can have unintended consequences. Every event is possibly a new event, and this means that although you can have insights from your past experiences, you can’t control the outcomes. You can’t simply copy and paste because the context in the current system is different from what you have experienced in the past.

4. Always get multiple perspectives (reality is multidimensional and constructed).
Get multiple perspectives. To quote the great American organizational theorist, Russell Ackoff, “Reality is multidimensional.” To add to this, it is also constructed. Multiple perspectives help us to understand things a little better and provide a new perspective that we were lacking. Systems are also constructed and can change how reality appears, depending on our perspective.

5. Go inside and outside the system.
We can’t try to understand a system by staying outside it all of the time. Conversely, we can’t understand a system by staying inside it all the time. Go to the gemba (the actual workplace) to grasp the situation to better understand what is going on. Come away from it to reflect. We can understand a system only in the context of the environment and the interactions going on.

6. Embrace variety.
Similar to item four, variety is your friend in a complex system. Variety leads to better interactions that will help with developing new patterns. Our environment is not homogenous; if everybody were the same, we’d lack the requisite variety to counter the variety present in our environment.

7. Aim for effectiveness, not efficiency.
In complex systems, we should aim for effectiveness. Here, the famous Toyota heuristic, “Go slow to go fast” is applicable. Because each event is novel, we can’t always aim for efficiency.

8. Use heuristics, not rules.
Heuristics are flexible while rules are rigid. Rules are based on past experiences and may lack the variety needed in a given context. Heuristics allow flexibility for agents to change tactics as needed.

9. Experiment frequently with safe-to-fail, small experiments.
As part of prodding the environment, we should engage in frequent and small, safe-to-fail experiments. This helps us improve our understanding.

10. Understand that complexity is always nonlinear, so keep an eye out for emerging patterns.
Complexity is nonlinear, and this means that a small change can have an unforeseen and large outcome. Thus, we should be alert for any emerging patterns and determine our next steps. Move toward what we have identified as “good,” and move away from what we have deemed “bad.” Patterns always emerge bottom-up. We may not be able to design the patterns, but we may be able to recognize them as they are developed and potentially influence them.

Final words

My column has taken a simple look at CAS. There are lot more attributes to CAS that are worth pursuing and learning. Complexity Explorer from Santa Fe institute is a great place to start.

I’ll finish with a quote from the retired U.S. four-star general, Stanley McChrystal, from his book, Team of Teams (Portfolio, 2015):
“The temptation to lead as a chess master, controlling each move of the organization, must give way to an approach as a gardener, enabling rather than directing. A gardening approach to leadership is anything but passive. The leader acts as an ‘eyes-on, hands-off’ enabler who creates and maintains an ecosystem in which the organization operates.”

Always keep on learning....

First published on Harish's Notebook blog.

About The Author

Harish Jose

Harish Jose has more than seven years experience in the medical device field. He is a graduate of the University of Missouri-Rolla, where he obtained a master’s degree in manufacturing engineering and published two articles. Harish is an ASQ member with multiple ASQ certifications, including Quality Engineer, Six Sigma Black Belt, and Reliability Engineer. He is a subject-matter expert in lean, data science, database programming, and industrial experiments, and publishes frequently on his blog Harish’s Notebook.