Design of optimal digital filter using a parallel genetic algorithm

This paper describes the use of a parallel genetic algorithm to design a direct form of a finite word length (FWL), finite impulse response (FIR) low pass digital filter. The results of the proposed design technique are compared to two other proposed methods, an integer programming technique and roundoff of the optimal floating point set of coefficients. Using the Chebyshev metric as the comparison criterion, design of the low pass filter by the parallel genetic algorithm technique proved to be superior to the other methodologies