
Velocity Adaptation in Particle Swarm Optimization
Sabine Helwig^{1},
Frank Neumann^{2}, and
Rolf Wanka^{1}
^{1}Department of Computer Science, University of ErlangenNuremberg, Germany
{sabine.helwig, rwanka}@informatik.unierlangen.de
^{2}MaxPlanckInstitut für Informatik, Saarbrücken, Germany
fne@mpiinf.mpg.de
Summary.
Swarm Intelligence methods have been shown to produce good results in
various problem domains. A wellknown method belonging to this kind of algorithms
is particle swarm optimization (PSO). In this chapter, we examine how adaptation
mechanisms can be used in PSO algorithms to better deal with continuous optimization
problems. In case of boundconstrained optimization problems, one has to cope with the
situation that particles may leave the feasible search space. To deal with such situations,
different bound handling methods were proposed in the literature, and it was observed
that the success of PSO algorithms highly depends on the chosen bound handling
method. We consider how velocity adaptation mechanisms can be used to cope with
bounded search spaces. Using this approach we show that the bound handling method
becomes less important for PSO algorithms and that using velocity adaptation leads
to better results for a wide range of benchmark functions.
In: B. K. Panigrahi, Y. Shi, M.H. Lim (eds.), Handbook of Swarm Intelligence  Concepts,
Principles and Applications, Springer, pp. 155173, 2011.
[doi:10.1007/9783642173905_7]
BibTex entry

