Ant systems, also known as ant algorithms or ant colony optimizers, are multi-agent systems in which the behavior of each computer agent is inspired by the behavior of real ants. When real-world ant colonies are given access to a food source that has multiple approach paths, most ants end up using the shortest and most efficient route. To expedite this process, some ant species deposit a chemical substance called pheromone on the ground when traveling from the nest to the food source. While the process iterates, pheromones are deposited at a higher rate on the shorter paths than the longer ones. When the other ants arrive at a decision point, like an intersection between various paths, they make a probabilistic choice based on the amount of pheromones they smell. After several trips, nearly all of the ants are using the shortest path due to the high concentration of pheromones deposited.

In a computer program, the ant algorithms simulate the deposition of pheromones along the various paths they travel upon. After several iterations, the optimal path can be identified by the high concentration of ants traveling on it. Ant algorithms are one of the best examples of swarm intelligence systems, and have been used to solve numerous optimization problems, ranging from the classical traveling salesman problem to routing in telecommunication networks.