Skip Navigation

Minority rule works for animals

February 2, 2005 By Roxanne Khamsi This article courtesy of Nature News.

Algorithm shows how honeybees and fish follow a lead.

Please log in to rate this page.

View Comments

A handful of clued-up individuals can steer a swarm of honeybees or a school of fish in the right direction, research suggests. The finding could help engineers to deploy robots more effectively in the future.

A computer algorithm has shown that animals with simple behaviour can use simple rules to make complex group decisions. And as the group becomes larger, the proportion of individuals who need to know what they are doing falls.

"I didn't expect any of these results with such a simple model," says Iain Couzin of the University of Oxford, UK, an animal-behaviour expert and one of the algorithm's designers. "I'm still surprised that the group is so good at collectively making decisions."

Wild idea

The algorithm gives its virtual animals several rules of thumb. One is that they try to avoid becoming cut off from the crowd. Wild animals try to avoid this too, says Couzin: a herring will die from stress if it is isolated from its school. Another rule is that group members should avoid getting so close that they crash into one another.

In addition to these opposing forces, the virtual animals are given a power of persuasion that depends on their desire to lead the group in a specific direction. Completely naive animals, with no idea of where to go, have zero power.

The simulations show that even when naive and informed individuals cannot recognize one another, the novices spontaneously respond to decisions by the experts, because they follow their tendency to stick with the group.

Moving along

More surprisingly, the computer simulations reveal that as a group grows larger, it requires a smaller percentage of leaders.

Researchers already know that a few individuals, perhaps 5% of the total, can guide a honeybee swarm. And the simulations show that a larger proportion isn't necessary, says Couzin: "This model explains how that kind of complex information transfer can occur without requiring individual cognitive complexity," he says. The results of the study appear this week in Nature1.

The findings could help engineers design protocols to guide groups of robots operating beyond human control, such as in outer space or the deep sea. Robots could use simple collective decision-making to travel wisely in dangerous environments, Couzin speculates.

To see how well the findings apply to animals with more sophisticated cognition, the researchers have begun to study human crowds. They hope it will help to explain how people behave during evacuation procedures.

"The fact that you don't need many individuals of the group to know where to go is interesting," says Jon Kerridge, an expert in pedestrian movement at Napier University in Edinburgh, UK. But, he cautions, the model may not apply directly to more complex animals. "Humans have the ability to communicate, and they use that skill."


  1. Couzin I. D., et al. Nature 433, 513 - 516 (2005).


User Tools [+] Expand

User Tools [-] Collapse

Pinterest button


Please log in to add this page to your favorites list.

Need Assistance?

If you need help or have a question please use the links below to help resolve your problem.