Model cooperation in Particle Swarm Optimization

This article explores cooperation of different mathematical models in the same system if the Particle Swarm Optimization (PSO) method is used for system parameter identification. Four different realistic system mathematical models of the second order were used to create different state spaces of the agents´ positions. All the experiments were evaluated in wide and narrow intervals around the fitness function global minimum with two different PSO method initial agents´ setup. Better results can be seen in some special cases when comparing model cooperation to single model optimization.