DocumentCode :
2254763
Title :
Ameliorated particle swarm optimization by integrating Taguchi methods
Author :
Liu, Chuan-Hsi ; Chen, Yen-liang ; Chen, Jen-Yang
Author_Institution :
Dept. of Mechatron. Technol., Nat. Taiwan Normal Univ., Taipei, Taiwan
Volume :
4
fYear :
2010
fDate :
11-14 July 2010
Firstpage :
1823
Lastpage :
1828
Abstract :
In this study, a novel particle swarm optimization (PSO) integrated with Taguchi method will be introduced. We use Taguchi method to assist PSO in finding the optimum in each dimension of position vectors during iterations, and exploit those optima to derive a new best-adaptive position vector (particle) afterward. Through verification over six benchmark functions, we have compared this PSO-Taguchi algorithm with the traditional global and local versions of PSO, and have found that the PSO-Taguchi method has a superior performance in convergence rate. In this paper, PSO will be first introduced. Then Taguchi method and its characteristics will be reviewed. Next, the issue of slow convergence speed with regard to the traditional PSO will be discussed. Finally, in order to solve this issue, a novel PSO-Taguchi algorithm will be proposed and verified through simulations.
Keywords :
Taguchi methods; particle swarm optimisation; PSO-Taguchi algorithm; Taguchi methods; ameliorated particle swarm optimization; best-adaptive position vector; Arrays; Benchmark testing; Convergence; Cybernetics; Machine learning; Optimization; Particle swarm optimization; Optimization technique; Particle swarm optimization (PSO); Taguchi method;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Machine Learning and Cybernetics (ICMLC), 2010 International Conference on
Conference_Location :
Qingdao
Print_ISBN :
978-1-4244-6526-2
Type :
conf
DOI :
10.1109/ICMLC.2010.5580960
Filename :
5580960
Link To Document :
بازگشت