Empirical computation of the quasi-optimal number of informants in particle swarm optimization

Abstract

In the standard particle swarm optimization (PSO), a new particle’s position is generated using two main informant elements: the best position the particle has found so far and the best performer among its neighbors. In fully informed PSO, each particle is influenced by all the remaining ones in the swarm, or by a series of neighbors structured in static topologies (ring, square, or clusters). In this paper, we generalize and analyze the number of informants that take part in the calculation of new particles. Our aim is to discover if a quasi-optimal number of informants exists for a given problem. The experimental results seem to suggest that 6 to 8 informants could provide our PSO with higher chances of success in continuous optimization for well-known benchmarks.