Application of a breeder genetic algorithm for finite impulse filter optimization

We describe in this paper the application of a breeder genetic algorithm to the problem of parameter identification for an adaptive finite impulse filter. This algorithm was needed due to the epistiasis phenomena, which is present for this type of adaptive filter. The results of the genetic algorithm were compared to the traditional statistical method and, we found that the breeder genetic algorithm was clearly superior in a multimodal space in most of the cases. However, the statistical least mean squares method is faster than the genetic algorithm. A hybrid method combining the advantages of both methods is proposed for real world applications.