Fu, Hua; Qiao, Dehao; Chi, Jihui A CIPSO-ENN coupling algorithm for nonlinear dynamic system parameter identification. (Chinese. English summary) Zbl 1249.93049 J. Xi’an Jiaotong Univ. 45, No. 2, 49-53 (2011). Summary: Aiming at the complexity, time varying and nonlinearity of the projects, a Chaotic Immune Particles Swarm Optimization — Elman Neural Network (CIPSO-ENN) coupling algorithm for identifying the parameters of nonlinear dynamic models is proposed, where the clonal selection of artificial immune system and chaotic mutation mechanism are embedded into standard particle swarm optimization. In the evolution of the particle swarm optimization population, this algorithm accelerates convergence of particle clonal selection and enhances the particle swarm local search capability after cloned particle chaotic mutation. Then CIPSO algorithm is merged with dynamic feedback Elman neural network to construct system identification models based on the CIPSO-ENN. The experiment results show that the identification model convergence rate is increased by 10 times and fitting accuracy is increased by 2 orders of magnitude compared with the pure Elman network identification method. MSC: 93B30 System identification 90C59 Approximation methods and heuristics in mathematical programming 90C10 Integer programming 93B40 Computational methods in systems theory (MSC2010) 92B20 Neural networks for/in biological studies, artificial life and related topics Keywords:chaos algorithm; clonal selection; Elman’s neural network; coupling algorithm; nonlinear dynamic system PDFBibTeX XMLCite \textit{H. Fu} et al., J. Xi'an Jiaotong Univ. 45, No. 2, 49--53 (2011; Zbl 1249.93049)