Swarm algorithms with chaotic jumps applied to noisy optimization problems

In this paper, we investigate the use of some well-known versions of particle swarm optimization (PSO): the canonical PSO with gbest model and lbest model with ring topology, the Bare bones PSO (BBPSO) and the fully informed particle swarm (FIPS) on noisy optimization problems. As far as we know, some of these versions like BBPSO and FIPS had not been previously applied to noisy functions yet. A hybrid approach which consists of the swarm algorithms combined with a jump strategy has been developed for static environments. Here, we focus on investigating the introduction of the jump strategy to the swarm algorithms now applied to noisy optimization problems. The hybrid approach is compared experimentally on different noisy benchmark functions. Simulation results indicate that the addition of the jump strategy to the swarm algorithms is beneficial in terms of robustness.