No abstract

Ant colony optimization (ACO) is a class of algorithms for tackling optimization problems that is inspired by the pheromone trail laying and following behavior of some ant species. In ACO, (artificial) ants construct candidate solutions to the problem instance under consideration. Their solution construction is stochastically biased by (artificial) pheromone trails, which are represented in the form of numerical information that is associated with appropriately defined solution components, and possibly by heuristic information based on the input data of the problem instance being solved. A key aspect of ACO algorithms is the use of a positive feedback loop implemented by iterative modifications of the artificial pheromone trails that are a function of the ants' search experience; the goal of this feedback loop is to bias the colony toward the most promising solutions. The ACO metaheuristic is a high‐level algorithmic framework for applying the above ideas to the approximate solution of discrete optimization problems. In this article, we review the most important developments on the algorithmic side of ACO and give a concise summary of approaches followed for the hybridization of ACO algorithms with other techniques, the main application areas of ACO, and the available theoretical results.

No abstract

No abstract