Digital filter design using genetic algorithm

Abstract

The authors present an algorithm for designing 1-D FIR filters using genetic algorithms. In addition, the authors examine the usefulness of various error norms such as I/sub 2/ and I/sub /spl infin// (minimax), and their impact on the convergence rate and the optimal result. The authors also present the application of various coding schemes applied to the filter coefficients and their effect in obtaining optimum filters. Examples of 1-D FIR filters are given using the above methodologies to illustrate the usefulness and utilities of the proposed techniques.