COSO: Community of Scientists Optimization

In this paper we propose COSO “Community Of Scientists Optimization”: a new evolutionary and population-based optimization algorithm using, as metaphor, the behavior of the research world, that is, probably, the most notable example of swarm intelligence. It is built as a variant of Particle Swarm Optimization (PSO) and the main idea introduced is, fundamentally, the use of a budget value associated to each population’s individual. Basing on these values, COSO is able to evolve the population size, generating and killing entities, other than the properties of the individuals. Experiments shown that COSO generally improves the convergence speed with respect to basic PSO