Artificial ants for automatic classification - Archive ouverte HAL Accéder directement au contenu
Chapitre D'ouvrage Année : 2010

Artificial ants for automatic classification

Résumé

Artificial ants have greatly contributed to the emergence of new optimization methods. However, the collective work of ants is not limited solely to the collection of food and other natural behaviors may yield applications in computer science. This chapter especially focuses on the automatic classification problem. Even if in terms of computer science, the search for an optimal classification relates to an optimization problem, new solving methods can be imagined in drawing inspiration from ants, but not limited to already known optimization capabilities. We will describe these solutions, beginning by succinctly reminding the classification issues. Then, we will present the two broad categories of artificial ant algorithms dealing with classification: the methods based on a grid, where ants move the data, as they would do in nature with their brood or waste (cemeteries), and the methods where each ant is associated with only one data to classify. Finally, this presentation is concluded with some specific applications of artificial ant algorithms, in the field of classification.
Fichier non déposé

Dates et versions

hal-01067324 , version 1 (23-09-2014)

Identifiants

  • HAL Id : hal-01067324 , version 1

Citer

Amira Hamdi, Violaine Antoine, Nicolas Monmarché, Adel Alimi, Mohamed Slimane. Artificial ants for automatic classification. Nicolas Monmarché, Frédéric Guinand and Patrick Siarry. Artificial Ants : from collective intelligence to real life optimization and beyond, Iste-Wiley, chap. 13, 2010. ⟨hal-01067324⟩
168 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More