A Study of Collapse in Bare Bones Particle Swarm Optimization

The dynamic update rule of particle swarm optimization is formulated as a second-order stochastic difference equation and general relations are derived for search focus, search spread, and swarm stability at stagnation. The relations are applied to three particular particle swarm optimization (PSO)...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on evolutionary computation Vol. 16; no. 3; pp. 354 - 372
Main Author: Blackwell, T.
Format: Journal Article
Language:English
Published: New York, NY IEEE 01.06.2012
Institute of Electrical and Electronics Engineers
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Subjects:
Algorithmics. Computability. Computer arithmetics
Applied sciences
Artificial intelligence
Bones
Collapse
Computational and artificial intelligence
Computer science; control theory; systems
Convergence
Dynamics
Equations
Exact sciences and technology
Mathematical model
Optimization
Particle swarm optimization
Random variables
Searching
Spreads
Stability
Stability analysis
Theoretical computing
Second order equation
Model matching
Probabilistic approach
Evolutionary algorithm
Updating
Computational and artificial intelligence
Swarm intelligence
Artificial intelligence
Adaptive method
Particle swarm optimization
ISSN:1089-778X, 1941-0026
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The dynamic update rule of particle swarm optimization is formulated as a second-order stochastic difference equation and general relations are derived for search focus, search spread, and swarm stability at stagnation. The relations are applied to three particular particle swarm optimization (PSO) implementations, the standard PSO of Clerc and Kennedy, a PSO with discrete recombination, and the Bare Bones swarm. The simplicity of the Bare Bones swarm facilitates theoretical analysis and a further no-collapse condition is derived. A series of experimental trials confirms that Bare Bones situated at the edge of collapse is comparable to other PSOs, and that performance can be still further improved with the use of an adaptive distribution. It is conjectured that, subject to spread, stability and no-collapse, there is a single encompassing particle swarm paradigm, and that an important aspect of parameter tuning within any particular manifestation is to remove any deleterious behavior that ensues from the dynamics.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ObjectType-Article-2
ObjectType-Feature-1
content type line 23
ISSN:1089-778X
1941-0026
DOI:10.1109/TEVC.2011.2136347