Abstract:
Several techniques have been proposed to extend the particle swarm
optimization (PSO) paradigm so that multiple optima can be located and
maintained within a convoluted search space. A significant number of these
implementations are subswarm-based, that is, portions of the swarm are
optimized separately. Niches are formed to contain these subswarms, a
process that often requires user-specified parameters. The vector-based
PSO uses a novel approach to locate and maintain niches by using
additional vector operations to determine niche boundaries. As the standard
PSO uses weighted vector combinations to update particle positions and
velocities, the proposed technique builds upon existing knowledge of the
particle swarm. Once niche boundaries are calculated, the swarm can be
organized into subswarms without prior knowledge of the number of
niches and their corresponding niche radii.
This paper presents the vector-based PSO with emphasis on its
underlying principles. Results for a number of functions with different
characteristics are reported and discussed. The performance of the
vector-based PSO is also compared to two other niching techniques for
particle swarm optimization.