A Study on IIR Filter Design Problem Using Real-Coded GA

Abstract

IIR filter design problem is nonlinear optimization problem and meta-strategy is often applied. The features of this problem are relative small number of design parameters, and all parameters take real value. In addition, there are interdependencies among design parameters. Keeping diversity is one of the key points to search good solutions in meta-strategy. GA (Genetic Algorithm) is well-known to provide such a search ability for a combinatorial optimization problem. It is impossible to directly apply GA to IIR filter design problem because GA basically treats discrete values. Although real-coded GA that can be applied to real values has been proposed so far, there are no examples of applying real-coded GA to IIR filter design problem. In this paper, the suitability of real-coded GA for IIR filter design problem is studied.