Genetic algorithm in the design of FIR filters

The author´s aim is to describe a method of designing finite impulse response (FIR) filters that is automatic, rapid, and gives filter realisations of near minimal computational complexity. Existing methods of filter design are reviewed to show that none possesses all these features. These methods include recent work using a sequential algorithm that produces realisations of guaranteed minimum complexity and thus provides a reference for the results described in the paper. Genetic algorithms are described, and a method of representing the problem of filter synthesis for solution by a genetic algorithm is given. Results are presented, demonstrating the suitability of the genetic algorithm design method