Application des algorithmes de colonies de fourmis pour l’optimisation et la classification des images

Abstract

Currently, the natural world has become the basic tool in scientific research. Because the invention is an imitation of a physical phenomenon, transformed to a mathematical formulation and become a mathematical model. In recent years, collective intelligence has attracted great interest most researchers in both biology and computer science. Biologists interested in insect societies seek to understand the mechanisms underlying the complex processes that lead to collective behavior research (foraging for survival of a colony, nest building, division of labor, ...) . It inspired the phenomena studied to develop new distributed and adaptive algorithms solving (pattern recognition, classification, optimization, routing in networks,...). The problems of optimization and classification are the basis of all vital operations and many methods have been developed to address the objectives of optimization and partitioning. Among these methods, ants colonies algorithms that form a class of metaheuristics recently proposed for these types of problems. We will apply these algorithms for optimization and classification of several types of data (satellite images and the base of Machine Learning Repository UCI and others) and see their contribution to solving these types of problems. keywords : Classification, optimization, clustering, ants colonies, Satellites images, Meta-heuristic, AntClust, AntClust adapted, ACOClust, 0, Machine Learning Repository