System identification of linear unstable plants using Craziness based Particle Swarm Optimization algorithm

This paper delivers an efficient soft computational approach towards the identification of a closed loop system with an unstable forward path system function. The identification is carried out with a modified Particle Swarm Optimization (PSO) algorithm called as Craziness based PSO (CRPSO). A parametric form of system model has been incorporated to identify the parameter vector associated with the system by minimizing the output Mean Square Error (MSE) between the actual and estimated model. The model accuracy and performance of CRPSO is discussed with the help of convergence of the output MSE along with the statistical information under the influence of uncertainties. The accuracy of estimation is also reported with the parameters´ convergences of the adaptive estimated model along with their maximum and minimum deviations.