Swarm Intelligence

When we look at the complex movements of a school of fish or a flock of birds, or the organisation involved in a bee hive or an ant colony, we might be convinced that fish, birds, bees and ants were particularly intelligent, particularly when we take their brain size into account. In actual fact the intelligence of these creatures is in their groups rather than in the individuals. Each individual is only responding using very simple rules and yet the behaviour of the whole flock, school, hive or colony is complex. This is called swarm intelligence. A whole school of fish swimming looks like an organism in its own right.

Another common characteristic of all these groups is that no one creature is in charge. The queen of a bee hive or ant colony lays eggs, but is in no way in control of the hive. The intelligent behaviour truly does emerge from the whole population.

Consider red ants. Each day they work out how many ants need to go out foraging and where they will go. They need to know if it is safe to go out and how many are needed around the colony. Ants communicate by touch and smell. Those from another colony smell different, as do those ants which work outside the nest.

Each morning patrollers are sent out and they do not return until they have found a food source. When the patrollers start arriving back they communicate with the foragers. When the foragers communicate with a number of patrollers, no less than about ten seconds apart, they follow the directions as displayed by a special dance to the food source. The patroller has also indicated how plentiful the food is in the dance. The decision to go and forage is simply made through these simple communications that only has minimal requirements on each ant.  Each ant only has information about what is happening in the immediate vicinity. No ant has the over all “big picture”.

As an ant finds a food source it releases a pheromone that other ants can recognize and follow. When they find the food source they also release the chemicals reinforcing the location as a source of food. The strength of the pheromone is thus a guide to other ants as to how much food is at the locations. The pheromone disperses over time, so when the food is all collected the pheromone trail stops and the ants search elsewhere for food.

When a hive of bees gets too large it divides into two Half the bees then go off in search of a new home. They leave and find a temporary place from which they send out scouts. In a similar way to the ants, they select the best suitable site for the new hive when enough scouts comeback in short succession.

This same swarm intelligence has real life applications in areas such as routing truck, scheduling airlines, telephone companies and guiding military robots. Electronic versions of a pheromone trail have helped trucks to receive goods from various depots where the prices of the goods are changeable more effectively than previously.

Telephone companies have used swarm intelligence based algorithms to choose the best way to route telephone calls, airlines use them to get to avoid waiting too long at the gates before take off.

The same principles have even been used in making business boards more efficient. They used the principles of identifying all possibilities when considering a problem, talking the ideas through for a while, then having a secret ballot to decide.

Whenever we get a large enough group of people get together, the way they move fits the patterns of swarms. Computer simulations of swarms can be useful for finding the most efficient way for people to get out of a building quickly, or highlight how design changes in the shape of a building might make egress more efficient.

Swarming behaviour has obvious survival advantages. It would be harder for a predator to attack a flock of birds or a school of fish that is in formation than it would to attack a single bird trying to escape. The large swarm can look like a large creature because of its organic like movements and this may also help ward off predators.

Swarms are therefore an effective strategy assisting creatures to survive with a minimum of organisation. Swarm dynamics can also be used to understand the spread of forest fires, the growth of species in an environment, where two or more swarms come together and compete in the environment.