Filter modeling using gravitational search algorithm

This paper is devoted to the presentation of a new linear and nonlinear filter modeling based on a gravitational search algorithm (GSA). To do this, unknown filter parameters are considered as a vector to be optimized. Examples of infinite impulse response (IIR) filter design, as well as rational nonlinear filter, are given. To verify the effectiveness of the proposed GSA based filter modeling, different sets of initial population with the presence of different measurable noises are given and tested in simulations. Genetic algorithm (GA) and particle swarm optimization (PSO) are also used to model the same examples and some simulation results are compared. Obtained results confirm the efficiency of the proposed method.