Application Of A Neural Network Based Genetic Algorithm To The Design Of Digital Filter

In this thesis, a BP neural network based GA (Genetic Algorithm) is proposed totake advantage of their complementary ability of local and global search for optimumsolutions. To show the effectiveness of this novel HGA (Hybrid GA), We haverespectively developed two application-oriented algorithm for the design of FIR andIIR digital filter that are widely used in a variety of signal processing problems.Several LP/HP FIR and IIR filters are designed based on the proposed HGA,also experimental simulations are performed to justify the effectiveness of the HGA.Simulation results demonstrate that as compared with corresponding results yieldedfrom other exiting design algorithms, the filters design by our HGAapproach are withbetter performance and less computation time.