An improved PSO algorithm encoding a priori information for nonlinear approximation

Authors:
Tong-Yue Gu;Shi-Guang Ju;Fei Han
Affiliations:
School of Computer Science and Telecommunication Engineering, Jiangsu University, Zhenjiang, Jiangsu, China;School of Computer Science and Telecommunication Engineering, Jiangsu University, Zhenjiang, Jiangsu, China;School of Computer Science and Telecommunication Engineering, Jiangsu University, Zhenjiang, Jiangsu, China
Venue:
ICIC'09 Proceedings of the Intelligent computing 5th international conference on Emerging intelligent computing technology and applications
Year:
2009

Citing 5
Cited 0

Encoding a priori information in feedforward networks

Neural Networks
Swarm intelligence

Swarm intelligence
Fundamentals of Computational Swarm Intelligence

Fundamentals of Computational Swarm Intelligence
Modified constrained learning algorithms incorporating additional functional constraints into neural networks

Information Sciences: an International Journal
A PSO Algorithm with the Improved Diversity for Feedforward Neural Networks

IITSI '09 Proceedings of the 2009 Second International Symposium on Intelligent Information Technology and Security Informatics

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper, an improved PSO algorithm for nonlinear approximation is proposed. The particle swarm optimization is easy to lose the diversity of the swarm and trap into the local minima. In order to resolve this problem, in the proposed algorithm, when the swarm loses its diversity, the current each particle and its historical optimium are interrupted by random function. Moreover, the a priori information obtained from the nonlinear approximation problem is encoded into the PSO. Hence, the proposed algorithm could not only improve the diversity of the swarm but also reduce the likelihood of the particles being trapped into local minima on the error surface. Finally, two real data in chemistry field are used to verify the efficiency and effectiveness of the proposed algorithm.