S

Swarm Behavior

Scientists frequently examine swarm behavior in animals. The seemingly chaotic behavior of bird flocks was often examined using state of the art GPS technology. The results of these studies provided insight into the intelligence of these swarms.

Swarm behavior researc

What is Swarming?

A basic definition is when several entities, usually animals, move together. However, it’s not always that simple and the reasons for this kind of swarm behavior are many and varied. What may look like chaos to outsiders is usually a very skilled and crafted situation that benefits those that are swarming.

Discover the variety of solutions offered by clickworker!

Animal Swarms

Usually, swarm behavior is seen in animals and especially in insects. The reasons for swarming can be related to food, migration, reproduction and in particular defence. There is power in numbers after all and smaller animals may find predators leave them alone if there are several of them. Starling birds flock together in huge swarms of up to one million to make it hard for larger birds to attack them. Curiously, bees swarm when their hive is too small, at this point they will look for a new hive and the swarm relocates.

Plant Swarms

Believe it or not, plants also swarm, even algae. Scientists have known about plant swarms since the 1800s. Roots are a great example of this, their patterns aren’t just random. Studies have shown that root pathways are based on communication to ensure optimum nutrition.

Human Swarms

Swarm behavior can often be observed in human beings. This behavior appears whenever strangers gather. The streams of people in a crowded shopping street organize themselves in such a way as to ensure that everyone finds their way. A direction of travel is automatically determined. However, if someone changes sides they become an obstacle because they are moving against the crowd. Interestingly, this phenomenon, known as flocking behaviour, is also used widely in mathematics and applied often in movies.

Robotic Swarms

Unsurprisingly, people have studied biological swarming behavior and applied it to robotics. This usually uses small, simple robots that do tasks in large numbers that couldn’t be done with just one unit. The use of sensors and programming sees the robots work together towards a common goal. Examples of this are drones used in military battles, and small robots used mining and search and rescue operations.

Interesting Swarm Facts!

  • Piranhas don’t swarm in scary packs. In fact, they are quite fearful and gather together to avoid being eaten.
  • Bacteria can swarm too. Some bacteria carry disease around the body in groups to maximise spread.
  • Roads and buildings have been designed based on the study of human swarm behaviour.
  • Lemmings do not gather in groups to jump off cliffs. The myth arose from occasional drowning when migrating en masse.
  • Human swarms are extremely predictable with weather being one of the largest influences.

Swarm Intelligence

What is Swarm Intelligence?

In 1989, two Electrical Engineering professors introduced the term “Swarm Intelligence” also known as SI. SI simply refers to the knowledge of a group or collective of animals, people, or things. In these groups or crowds all individuals are independent of one another, they don’t necessarily work with one another. Therefore, each entity provides their own solution to a problem or task separately from everyone or everything else. SI is used when a solution to a problem requires more than just one person’s answer or when more than one entity is required to complete a task. Interestingly, it has some similarities to crowdworking!

Swarm Intelligence in Action

Overall, most people know that knowledge gained from the collective is usually better than knowledge from an individual. The reasons for this are many, a single person should not be relied upon to know the answer to all questions or situations. They may be wrong, lack knowledge, make errors, only know a part answer, etc. To explain this imagine a container filled with 500 small balls. When one person is asked to guess how many are in there they say 400, which is a pretty good guess! However, if you ask 10 people to guess the average of their responses (adding all replies and dividing by 10) comes to 470 which is a far better guess! Such is the power of the crowd.

Swarm Behavior in AI

SI or AI?

In some cases, it may seem like SI is used the same as AI but they’re actually different. AI is mainly based on learning via statistics or logical reasoning. Whereas SI is based on the behavior of a group of entities. Of course, there can be some overlap and in many situations, both can be used at the same time. However, SI can influence and assist AI in several ways.

Artificial Swarm Intelligence

Also known as ASI, artificial swarm intelligence is a way of boosting intelligence gained from collective groups of humans using algorithms. These algorithms come from analyzing swarm behavior in the animal kingdom such as ants, birds, and bees. Understandably, this is why it’s also sometimes called Human Swarming. This interactive technology brings together networks of humans. When working on a task they will work individually but as a group, this is then studied and analized. With the addition of AI there is an increase in much better decisions being made. Whilst another human in the group doesn’t tell someone else what to do, they may influence the decision of the other person.

Examples of SI Applications

Telecoms

Originally created by engineers at Hewlett Packard, British Telecom (BT) and many other telecom providers use swarm intelligence to assist their call centre staff and maximise the performance of their internet traffic. This was done by creating something called “digital ants” because of the way the algorithm is based on the swarm behavior of ants. These digital ants hypothetically run along the free phone lines to enable staff to route calls with ease. They can also reroute internet traffic to less congested routes resulting in faster internet speeds for their customers.

Recruitment

Companies can spend a lot of time trying to hire the right person for a job. Sifting through resumes and processing interviews can be a long and arduous task and may not result in the right person for the position. Going back to ant behaviour again, software has been developed to assist this situation. Similarly, swarm behavior in bees has helped, bees do a dance to signal a new hive or source of food to others. When a company is wanting to hire people for the lowest pay, they can be notified of those who will accept that kind of pay. Thus the right person for the job is linked to the role.

Manufacturing

Thinking about bees again, we know that they work together towards a common goal. So do factories, everyone and all the machinery works towards the production of specific items. The way bees allocate work is interesting, when workload increases, extra bees are sent to assist in getting the task done. A car factory utilized this in regard to the demand of certain paints. This created a self-governing system that sourced a specific paint color from places that had an abundance instead of relying on a usual supplier or the cheapest or nearest that may have run out.