Selfish herd optimization algorithm based on chaotic strategy for adaptive IIR system identification problem

• Published in: Soft computing (Berlin, Germany) Vol. 24; no. 10; pp. 7637 - 7684
• Main Authors: Zhao, Ruxin, Wang, Yongli, Liu, Chang, Hu, Peng, Jelodar, Hamed, Yuan, Chi, Li, YanChao, Masood, Isma, Rabbani, Mahdi, Li, Hao, Li, Bo
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.05.2020; Springer Nature B.V
• Subjects: Adaptive algorithms; Adaptive systems; Artificial Intelligence; Computational Intelligence; Control; Engineering; Evolutionary computation; Heuristic methods; Identification systems; IIR filters; Impulse response; Iterative methods; Laplace transforms; Local optimization; Mathematical Logic and Foundations; Mathematical models; Mean square errors; Mechatronics; Methodologies and Application; Methods; Optimization algorithms; Parameter estimation; Parameters; Particle swarm optimization; Researchers; Robotics; Search algorithms; Search methods; Signal processing; System identification; Chaotic strategy; IIR system identification problem; Local optimization strategy; Adaptive algorithms; Selfish herd optimization algorithm
• ISSN: 1432-7643; 1433-7479
• DOI: 10.1007/s00500-019-04390-9

The design method of adaptive infinite impulse response (IIR) filter is a challenging problem. Its design principle is to determine the filter parameters by the iteration process of the adaptive algorithm, which is to obtain an optimal model for unknown plant based on minimizing mean square error (MSE). However, many adaptive algorithms cannot adjust the parameters of IIR filter to the minimum MSE. Therefore, a more efficient adaptive optimization algorithm is required to adjust the parameters of IIR filter. In this paper, we propose a selfish herd optimization algorithm based on chaotic strategy (CSHO) and apply it to solving IIR system identification problem. In CSHO, we add a chaotic search strategy, which is a better local optimization strategy. Its function is to search for better candidate solutions around the global optimal solution, which makes the local search of the algorithm more precise and finds out potential global optimal solutions. We use solving IIR system identification problem to verify the effectiveness of CSHO. Ten typical IIR filter models with the same order and reduced order are selected for experiments. The experimental results of CSHO compare with those of bat algorithm (BA), cellular particle swarm optimization and differential evolution (CPSO-DE), firefly algorithm (FFA), hybrid particle swarm optimization and gravitational search algorithm (HPSO-GSA), improved particle swarm optimization (IPSO) and opposition-based harmony search algorithm (OHS), respectively. The experimental results show that CSHO has better optimization accuracy, convergence speed and stability in solving most of the IIR system identification problems. At the same time, it also obtains better optimization parameters and achieves smaller difference between actual output and expected output in test samples.