DocumentCode :
3098266
Title :
Simple paricle swarm optimization
Author :
Chen, Chang-Huang ; Sheu, Jia-shing
Author_Institution :
Dept. of Electr. Eng., Tungnan Univ., Taipei, Taiwan
Volume :
1
fYear :
2009
fDate :
12-15 July 2009
Firstpage :
460
Lastpage :
466
Abstract :
It is well known that the dynamic properties of the particles in the particle swarm optimization (PSO) can be described by a second-order difference equation. The convergent properties of the particle are then governed by the roots of the characteristic equation. The roots, or referred to eigenvalues, are functions of the coefficients, which are determined by the inertia weight and acceleration constants of PSO. Inspecting the characteristic equation, it is found that using less parameter for PSO is possible. Two versions of simplified PSO are thus derived directly from the characteristic equation. By testing on a set of benchmark functions, the feasibility and effectiveness of the proposed algorithm are validated. The experimental results demonstrate good performance, especially in multimodal functions, compared with the classical PSO. A byproduct of saving computational operations is also achieved with fewer parameters.
Keywords :
difference equations; eigenvalues and eigenfunctions; particle swarm optimisation; PSO; acceleration constant; characteristic equation root; eigenvalues; inertia weight; multimodal function; particle swarm optimization; second-order difference equation; Acceleration; Benchmark testing; Cybernetics; Difference equations; Eigenvalues and eigenfunctions; Machine learning; Machine learning algorithms; Mathematical model; Optimization methods; Particle swarm optimization; Evolutionay algorithms; Particle swarm optimization; Swarm intelligence;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Machine Learning and Cybernetics, 2009 International Conference on
Conference_Location :
Baoding
Print_ISBN :
978-1-4244-3702-3
Electronic_ISBN :
978-1-4244-3703-0
Type :
conf
DOI :
10.1109/ICMLC.2009.5212546
Filename :
5212546
Link To Document :
بازگشت