Application Of Genetic Algorithms In The Design Of Digital Filters

A modified genetic algorithm (GA) is proposed and its applications are studied in this thesis. Applications of genetic algorithm with guiding strategy in the design of digital filter (FIR DF, IIR DF) are also addressed.In chapter 1, Basics of GA, digital filters and the definition of digital filter parameters are briefly reviewed.In chapter 2, the principle, design method, technology requirements of digital filters (FIR DF, IIR DF) are presented. In chapter 3, the history and principle of the original GA and modified Gas together with their applications are briefly surveyed. A novel modified GA proposed by the author is addressed and the difference between traditional GA and our modified GA is given in detailed.In chapter 4, Applications of our modified GA with guiding strategy in the design of FIR and IIR filters are presented. The important concept of frequency sampling points and its weight coefficients are introduced. The Weight Least Mean Square criterion between the desired frequency response and corresponding filter response is defined as fitness function of GA, and corresponding representations are derived. Some FIR and IIR filter design examples are given. Computer simulations are conducted, and the performance improvements of the FIR and IIR designed by our modified GA are demonstrated.In chapter 5, Concluding remarks on our contribution presented in the thesis are given and the future works are also proposed.