Dynamic particle swarm optimization via ring topologies

ABSTRACT

Particle Swarm Optimization (PSO) has been proven to be a fast and effective search algorithm capable of solving complex and varied problems. To date numerous swarm topologies have been proposed and investigated as a means of increasing the effectiveness of the generalized algorithm. Typical topologies employ static arrangements of particles defined at the beginning of execution and remaining constant throughout run-time. Topologies that do allow for restructuring, often do so according to predefined rules that limit the opportunity and manner in which the topology can change. Recent investigations have shown that dynamically redefining a topology by stochastically re-organizing the swarm at periodic intervals improves performance for certain types of problems. In this work the effectiveness of a novel topology "Dynamic Ring" and a derivative of the {}"Dynamic Multi Swarm PSO" topology dubbed "Dynamic Multi Swarm with Ring" are investigated. We show that these two new topologies show generally enhanced performance relative to previously proposed topologies on a suite of twelve test functions.

REFERENCES

• 1: J. Kennedy, and R. Eberhart, Particle swarm optimization, in Proc. of the IEEE Int. Conf. on Neural Networks, Piscataway, NJ pp. 1942--1948, 1995
• 2: J. Kennedy , R. Mendes, Population structure and particle swarm performance, Proceedings of the Evolutionary Computation on 2002. CEC '02. Proceedings of the 2002 Congress, p.1671-1676, May 12-17, 2002
• 3: J.J. Liang, P.N. Suganthan: Dynamic Multi-Swarm Particle Swarm Optimizer. Swarm Intelligence Symposium, 2005. Proceedings 2005 IEEE. pp. 124--129
• 4: J.J. Liang, P.N. Suganthan: Dynamic Multi-Swarm Particle Swarm Optimizer with Local Search. The 2005 IEEE Congress on Evolutionary Computation, 2005, Volume: 1, pp: 522--528

INDEX TERMS

Primary Classification:
  G. Mathematics of Computing
  G.1 NUMERICAL ANALYSIS
      G.1.6 Optimization

Additional Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.2 Automatic Programming

General Terms:
Algorithms

Keywords:
dynamic neighborhoods, particle swarm optimization