Constrained genetic algorithm design of finite precision FIR linear phase raised cosine filters

This paper investigates the use of a genetic algorithm in the design of finite wordlength linear phase raised cosine FIR filters. The design of such filters not only involves meeting a specific amplitude response but also imposes time domain constraints on the impulse response in order to ensure zero intersymbol interference (ISI). It is shown that a genetic algorithm provides simple and effective techniques for the design of finite precision filters with both frequency and time domain constraints. Furthermore, the algorithm computational time is low with respect to filter order and is independent of word length. Two filter design examples are included demonstrating the effectiveness of the design method