Bacterial foraging based identification of nonlinear dynamic system

Abstract

Identification of nonlinear dynamic system plays an important role in many applications such as control engineering, telecommunication and intelligent instrumentation. The present paper investigates on the use of Bacterial Foraging in identification of nonlinear dynamic systems employing an efficient Functional link artificial neural network (FLANN) model. The BFO is a derivative free optimization tool and hence does not permit the solution of connecting weights to fall in local minima. This potential tool is employed in the paper to update the weights of the FLANN model. To assess the performance of the new model simulation studies of both the BFO-FLANN and multilayered ANN (MLANN) identification models are carried out. These experiments reveal that the two models exhibit identical identification performance. But, the proposed model offers low computational complexity and achieves faster convergence compared to its MLANN counterpart.