Allocation of local and global search capabilities of particle in canonical PSO

ABSTRACT

This paper analyzes theoretically the exact sampling distribution of the particle swarm optimization (PSO) without any assumption imposed by all current analyses. The distribution of particles in the PSO in one-step transition is analyzed in details. Especially, local and global search capabilities of particles in the PSO are defined implicitly, and are allocated adaptively to show how the PSO works by several experiments. In essence, the PSO works by just allocating each particle in the swarm to finish two jobs in probability, locally searching the range around the current best positions and globally searching whole solution space. According to our definitions and analyses, the exact probabilities of the two jobs can be measured by theoretic derivations and experiments whilst how the PSO allocate the particles' search capabilities can be recognized clearly. So we can look into the PSO in depth.

REFERENCES

• 1: J. Kennedy and R. Eberhart, Particle Swarm Optimization, Proceedings of the IEEE International Conference on Neural Networks, 1995, pp. 1942--1948.
• 2: R.C.Eberhart and J.Kennedy, A new optimizer using particle swarm theory, Proc. of the 6th Int. Symp. Mcro Machine Human Science, 1995,39--43.
• 3: Riccardo Poli , David Broomhead, Exact analysis of the sampling distribution for the canonical particle swarm optimiser and its convergence during stagnation, Proceedings of the 9th annual conference on Genetic and evolutionary computation, July 07-11, 2007, London, England [doi>10.1145/1276958.1276977]
• 4: Riccardo Poli, On the moments of the sampling distribution of particle swarm optimisers, Proceedings of the 2007 GECCO conference companion on Genetic and evolutionary computation, July 07-11, 2007, London, United Kingdom [doi>10.1145/1274000.1274075]

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.8 Problem Solving, Control Methods, and Search

Additional Classification:
  J. Computer Applications
  J.4 SOCIAL AND BEHAVIORAL SCIENCES

General Terms:
Performance, Theory

Keywords:
allocation, global search, guide, local search, particle swarm optimization (PSO), sampling distribution