Particle Swarm Optimization on Manifolds
Tianlai Li · 2009
Particle Swarm Optimization(PSO) is an increasingly used optimization technique which is derived from swarm intelligence.Generally,all particles are assumed in a linear space that is uniformly distributed.The term manifold comes from geometry.Roughly speaking,a manifold is nearly a curved space.The purpose of this paper is to present a particle swarm optimization framework on manifolds,that is,on curved space.The newly proposed MPSO will apply for nonlinear,nonuniform data distribution.Convergence analysis was presented and performance of the new algorithm was examined with experimental data given.