Ants live in colonies that can contain millions of , yet no single ant is in . The queen lays eggs does not give orders. , ants communicate mainly chemicals called pheromones. When a worker finds , it leaves a scent trail on its way to the nest. Other ants the trail, and if they also find food, they it. In this way, the colony good routes without any planning.

In the early 1990s, the Italian computer scientist Marco Dorigo noticed that the way ants find short paths could help solve difficult problems in computing. He developed a method known as ant colony optimization.

The method is based on a simple observation. Suppose ants can reach a food source by two routes, one shorter than the other. Ants using the shorter route return sooner, so pheromone builds up on it more quickly. As more ants choose the stronger trail, the shorter route becomes increasingly popular. Meanwhile, pheromone slowly evaporates, so trails that are no longer used fade away.

In computer versions, virtual ants explore possible solutions to a problem, such as the best order in which a delivery truck should visit fifty stops. Good solutions are marked with stronger virtual pheromone, guiding later searches toward them.

Such methods do not always find the perfect answer, but they can find very good answers to problems that would take ordinary computers far too long to solve by checking every possibility. They have been used to plan delivery routes, create schedules, and direct data through communication networks.