Adaptive PSO for Online Identification of Time-Varying Systems

Abstract

Novel adaptive PSO (particle swarm optimization) algorithms called OPSO (online PSO) and εPSO are proposed. These algorithms are designed to prevent increasing of calculation cost in order to embed into online systems, and they have high adaptive performances to the environmental changes. Then, the efficiency of the proposed PSO algorithms are demonstrated by numerical simulation and the online identification of an actual dynamic system.