A modified particle swarm optimizer using an adaptive dynamic weight scheme

Abstract

Particle swarm optimization (PSO) is a stochastic, population-based optimization technique that is inspired by the emigrant behavior of a flock of birds searching for food. In this paper, a nonlinear function of decreasing inertia weight that adapts to current performance of PSO search is presented. Meanwhile, a dynamic mechanism to adjust decrease rates is also suggested. Through the experimental study, the new PSO algorithm with adaptive dynamic weight scheme is compared to the exiting models in terms of various benchmark functions. The computational experience shows some great promise. © Springer-Verlag Berlin Heidelberg 2007.