Swarming Starlings

Starlings (and other birds, ants, fish, and termites) follow simple rules to coordinate group behaviour without a boss.

You have probably seen birds swarming in the sky. Did you ever wonder how they know where to go? Who is the boss? Birds such as the European Starling form flocks consisting of thousands of individuals moving in a seemingly coordinated way. Ants form highways of individuals, finding and foraging on food sources far away from their nests. Other insects, such as termites, build skyscraper-like nests that regulate internal temperature and are home to millions of individuals. All these behaviours are accomplished with no overseeing director, there is no single individual responsible for coordinating the others. How do they do it?

Take your group outside (indoors is also possible, you need enough room to move around) for a demonstration of the power of self-organisation through individuals following a simple set of rules. The instructions for the activity below are adapted from ICOSYSTEM. You can do this exercise with 15 or more (up to 150) pupils.

First set the context: ask everyone to stand in a circle. Explain to the group that this activity consists of three rounds and that you will pause after each round to debrief. Also, tell the group that all three rounds are done without talking and indicate the boundaries of the space in which you will do the exercise (where can they go).

• Part 1: First ask everyone to randomly select two individuals – person A and person B, without indicating who they select. Person A is a predator and person B is a shield. Ask the participants to move so that they always keep B in between themselves and A – so that B protects them from A. Everyone will mill about in a seemingly random fashion and will soon begin to ask why they are doing this. When this happens, ask them to stop.
• Check in with participants: what were you doing? What do you think, could this have gone on forever? What if there would have been no boundaries to the space?
• Part 2: Now ask them to come back in a circle and select two individuals without letting them know. Again, there is a person A and B. This time, they are the protector themselves so they have to move in such a way that they keep themselves in between A and B at all times. The results are striking. Almost instantaneously the whole room will implode on itself with everyone clustering in a tight knot.
• Check in with participants: what happened here, why? How does the change in just one change the organisation of the players?
• Part 3: Finally ask everyone to come back into the circle. Again, randomly select two individuals – person A and person B, without indicating who they select. Next, ask the participants to position themselves in such a way that they have equal distance to A and B. Participants need to keep moving until everybody succeeds. At some point (surprisingly quick even with groups over 60 people) an equilibrium is reached.
• Check in with participants: what happened here, why? What would have happened if we assigned one person to create this system?

Simple explanation

Just as in the exercise, flocks of birds use few simple rules. Each bird follows three simple rules. The rules are 1) get as close as you can, 2) keep a (certain) distance, 3) go in the same direction as your neighbour birds. Rules 1 and 2 are ‘balancing’ rules. Without rule #2 the birds might collapse into each other and without # 1 they may get too far from each other. Next to having the rules, and all birds understanding them, it is also important that each bird can always see if they adhere to the rules.

In the second round of the exercise only one rule was changed. But it had a great impact. Because everybody now needs to be in the middle (between A and B), nobody can be on the outside, and that is why everybody will get together.

Starlings can manage these high-performance acrobatics by using a couple of simple rules. They pay close attention to the speed and direction of the other starlings around them. However, starlings don’t pay attention to all the other birds in the flock at once. A starling just needs to pay attention to the movements of the seven other starlings closest to it.

More detailed explanation

If Starlings pay attention to any fewer than seven neighbours, then there’s not enough reliable information for a starling to maintain flying with precision within the flock. Any more, and there’s too much information to process quickly and make real-time decisions. It turns out that seven neighbours is the ideal number, regardless of how large or dense the flock is. The simple rules they use are related to distance (not too far away from the seven other starlings, but also not too close to them - to avoid collisions) and to direction (head in the same direction as the seven others close by).

Self-organised biological systems work with positive and negative feedback loops. Positive feedback can be explained as ‘keep doing’ and negative feedback as ‘stop doing’. Feedback loops often work in concert with one another. Feedback systems manifest in a set of rules that an individual follows. For example, swarming can be modelled by giving individuals three rules to follow, or two positive feedback loops and one negative. The first is cohesion, a positive feedback loop where individuals are attracted to one another. The larger the group or the closer the individual the greater the attraction is. The second, separation, is a negative feedback loop that limits the first. While the individual birds are attracted to one another, within certain proximity they are repulsed. The birds therefore have a preferred amount of minimum distance between one another. Finally, the last rule is alignment, a positive feedback loop. The birds modify their heading to match the heading of their neighbour.

When members of a group decide which rules are important to them, and everybody is committed to following them, much can be achieved in a shorter time, with less energy and sometimes less materials. For example, in the northern part of the Netherlands, there are towns that removed all traffic signs and created ‘shared spaces.’ There is no specific road or pedestrian area or cycling lane; all traffic users can go anywhere. At first sight it is more chaotic and therefore people are more cautious and pay more attention resulting in less accidents.

Questions for discussion: “What rules do you think might emerge from this situation? Where else might this approach help…somewhere in the school?”

There are now robots that work together by using simple rules. These robots can be programmed using algorithms inspired by swarming. These autonomous artificial swarms of robots have potential uses for search and rescue missions, construction efforts, environmental remediation, and medical applications.

Another example of self-organisation is the ‘Adopt-a-fire-hydrant’ app. With this app, citizens (in cities that get lots of snow in wintertime) can adopt a fire hydrant and commit themselves to keep the fire hydrant clean. In most cities, it is the municipality/fire brigade that needs to keep the fire hydrants free from snow. This means they have to drive around, see if they are snow-free, and if not start cleaning them. As you can imagine, this ‘centralised’ way of organising costs a lot of time and money. By having citizens adopt a hydrant and keeping it clean, it becomes a decentralised and self-organised activity, saving the fire department and local government a lot of money and time. Now they don’t have to drive around and monitor if they are snow-free and ready to use. People are willing to keep the hydrants snow-free as they depend on them in case of a fire.

Important to note is that self-organisation is not complete freedom. Having some rules and boundaries creates safety resulting in more freedom to act. See this video about the famous playground experiment.

Ask pupils what rules there are in their group. How many are there? Who makes the rules? Who ensures the rules are followed? What are some simple rules they could use and govern through self-organisation, so that they may skip some rules? What simple rules could your class come up with to make things go smoother?

Some STEAM opportunities include:

• Use of logical reasoning to predict simple behaviours.
• Asking questions and making observations.
• Analyse advantage and disadvantage of different adaptations/behaviours.
• Apply learning to real world problems.
• Making predictions.

Take a look at this video of swarming starlings (click here).

Read how scientists are creating swarming roborts (find out more).

Watch how termites are inspiring a robot construction team (click here).