Principal component particle swarm optimization (PCPSO)

Author

Voss, Mark S.

Author_Institution

Prediction Eng. Co., Willoughby Hills, OH, USA

fYear

2005

fDate

8-10 June 2005

Firstpage

401

Lastpage

404

Abstract

Particle swarm optimization (PSO) is based on the notion of particles flying through solution space. Each particle is assumed to have n-dimensions that are mapped to the variables of the function that is being evaluated. The standard PSO algorithm updates a particle by moving towards the particle´s past personal best and the best particle that has been found. This paper introduces the principal component particle swarm optimization (PCPSO) procedure. The principal component particle swarm optimization procedure flies the particles in two separates spaces at the same time; the traditional n-dimensional x space and a rotated w-dimensional z space where m ≤ n. The Griewank function is used for introducing the PCPSO algorithm and a PCPSO time complexity study.

Keywords

computational complexity; particle swarm optimisation; principal component analysis; Griewank function; PCPSO algorithm; PSO algorithm; particle swarm optimization; principal component particle swarm optimization; time complexity; Covariance matrix; Eigenvalues and eigenfunctions; Equations; Particle swarm optimization; Principal component analysis; Vectors;

fLanguage

English

Publisher

ieee

Conference_Titel

Swarm Intelligence Symposium, 2005. SIS 2005. Proceedings 2005 IEEE

Print_ISBN

0-7803-8916-6

Type

conf

DOI

10.1109/SIS.2005.1501651

Filename

1501651