1. HOW CONSTRAINED SPACE AFFECTS THE RESULTS IN THE PARTICLE SWARM OPTIMIZATION.
- Author
-
Nemeckova, Michala
- Subjects
PARTICLE swarm optimization ,MATHEMATICAL optimization ,ALGORITHMS ,ARTIFICIAL neural networks ,SWARM intelligence - Abstract
Particle swarm optimization (PSO) is a stochastic computational technique for finding optimal regions of complex search spaces. The method was inspired by social behavior of organism and it belongs to the evolutionary computation (EC) to the group of swarm intelligence (SI). It has a few parameters to adjust and the method is easy to implement, which are the main advantages. The main aim of this paper is to find out, if limitation of the search space influences the optimization process. In this paper, we compare two modifications of PSO algorithm - variant using the parameter of inertia weight and variant with constriction factor. Both variants are solved separately for constrained and unconstrained space. Constrained area is when the boundaries are clear district, unconstrained is when particles can cross the boundaries. For comparison, eleven benchmark functions which were prepared for the special session on real-parameter optimization of CEC 2005 were used. Differences in results between constrained and unconstrained area is presented in this paper. It was found, that constrained area of search space gives better results. In additional, variant of PSO algorithm with linearly decreasing inertia weigh is better for chosen optimization problem. [ABSTRACT FROM AUTHOR]
- Published
- 2013