Title: A novel particle swarm algorithm for solving parameter identification problems on graphics hardware
Authors: Jing Wang, Zhijian Wu, Hui Wang
Addresses: State Key Lab of Software Engineering, Wuhan University, Wuhan, Hubei, 430072, China; School of Software and Communication Engineering, Jiangxi University of Finance and Economics, Nanchang, Jiangxi, 330013, China. ' State Key Lab of Software Engineering, Wuhan University, Wuhan, Hubei, 430072, China. ' State Key Lab of Software Engineering, Wuhan University, Wuhan, Hubei, 430072, China
Abstract: This paper presents a fine-grained novel particle swarm optimisation (PSO) algorithm on graphics hardware. It has good performance on a collection of parameter identification problems. In this algorithm, a generalised opposition-based learning (GOBL) strategy is embedded into PSO algorithm. This strategy can transform the current solution space to provide more chances of finding better solutions, and the parallel computing on graphics hardware can accelerate the convergence rate significantly. Experiment results show that the novel algorithm on graphics hardware has not only a good tolerability with the noise in the observed data but also a very high speedup.
Keywords: particle swarm optimisation; PSO; graphics hardware; opposition-based learning; OBL; parameter identification; inverse problem; parallel computing.
International Journal of Computational Science and Engineering, 2011 Vol.6 No.1/2, pp.43 - 51
Published online: 13 Jul 2011 *Full-text access for editors Access for subscribers Purchase this article Comment on this article