Particle swarm optimization with turbulence (PSOT) applied to thermal-vacuum modelling

Particle swarm optimization with turbulence (PSOT) is, in this paper, applied to find out fuzzy models to represent dynamic behavior of space systems that lie underneath the space qualification process. In optimization area, each minimal improvement in results may represents a maximal, precious meaning and PSOT improve the performance of the established particle swarm optimization (PSO) by introducing a slight variation, which simulates the action of an atmosphere turbulence to escape from local minima. This paper trades off the results of original PSO presented in a previous paper and PSOT both intertwined with Takagi-Sugeno (TS) fuzzy modeling dealing with experimental results of a thermal-vacuum system. Particle swarm optimization with turbulence has demonstrated to be a good alternative by taking into account the velocity of convergence to better solution and the total optimization time in generating dynamical models to the proposed system.

I. Introduction

Efforts have been dispensed to improve the capacity of modeling dynamic systems. With high complexity characterizing many of these systems, techniques of nonlinear modeling are used to approximate a model to the given dynamic system.