Particle swarm optimization with approximate gradient
Naoya Nakagawa, Atsushi Ishigame, Keiichiro Yasuda · IEEJ Transactions on Electrical and Electronic Engineering · 2008
Abstract This paper presents a new Particle Swarm Optimization (PSO) technique that uses the approximate gradient of the object function. Sensitivity is not normally used in the PSO algorithm, so it is expected that the present approach employing the approximate gradient can provide a more efficient search. The main advantage of the present PSO approach is that, in contrast to approaches using the gradient, by using the approximate gradient it is also possible to apply PSO to nondifferentiable functions. In this paper, we shall verify the effectiveness of the proposed method through its application in several benchmark problems. © 2008 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.