Une méthode inspirée de comportements coopératifs observés dans la nature : l'optimisation par essaim particulaire

First we present in this paper the general framework of "metaheuristics'; which are numerical algorithms emerged in the early 1980s with a common ambition: to solve at best hard optimization problems. Then we present the Particle Swarm Optimization. It is a technique based on the concept of cooperation between agents (particles) that can be seen as animals having rather limited capacity (low memory and reasoning faculties). The exchange of information between them makes that they nevertheless come to solve difficult problems, as is the case, for example, of bees living in swarm (exploitation of food sources, and so on).