An Improved Identification Method for Multivariable System

Abstract

Due to many classical identification methods cannot be directly used for closed-loop control system, an improved identification method is proposed to simultaneously identify model parameters and the structure. The improved identification method used the genetic algorithm to estimate the initial search scope for the PSO algorithm, and then used the search result as the initial value of the Rosenbrock algorithm. On the basis the genetic algorithm to estimate is introduced to provide the rough initial search scope for the presented algorithm to improve the validity and accuracy. Simulation results show that compare with the PSO algorithm, the inertia weight variation PSO algorithm and the PSO-SQP algorithm proposed by Qibing Jin et al, the presented algorithm improves the optimizing efficiency of the particle swarm.

Partially coupled extended stochastic gradient algorithm for nonlinear multivariable output error moving average systems

“…Most industrial processes are multivariable systems in essence (Jin et al , 2014). The choice of model representations plays an important role in the identification of multivariable systems.…”