Ant Colony Optimization

Scientists have always been fascinated by the problem solving capabilities of Nature: the evolution of the species, the nest building activity of termites, the way crystal forms out of shapeless materials are only a few of the possible examples. For long, biologists, physicists, and chemists have studied and tried to understand these phenomena, building explicative models. More recently, also computer scientists have become interested in these phenomena: a wide set of nature inspired optimization algorithms have been developed: among others, we may mention genetic algorithms, neural networks and simulated annealing. A recent and important class of nature inspired algorithms is that of ant algorithms [1]. These are algorithms inspired by the observation of social insect behavior, and in particular by the behavior of ant colonies. In these algorithms, the traditional emphasis on control, preprogramming, and centralization is replaced by an emphasis on autonomy, emergence, and distributed functioning. A particularly successful research direction in ant algorithms, known as Ant Colony Optimization [3,5], is concerned with applications to discrete optimization problems. The aim of this short note is to introduce and briefly describe origins and basic principles of Ant Colony optimization (ACO).