Improved Genetic Algorithm Based FIR Filter Design

The design of FIR filter is essentially an optimization problem. Genetic algorithms (GAs) are used to design the linear phase FIR filter with low and median taps. Improved crossover and mutation operators are proposed depending on the characteristic of the problem. The filter is designed using direct method when the tap is low (less than 18). Incremental tap method is quite useful in the design of median-tap filter (between 18-34). Both the two methods get good results. Comparisons with traditional design methods show that genetic algorithms based FIR filter design method has obvious advantage in small problem and is the powerful competitor in medium problem. Suggestions for expanding the adaptive range of GAs are proposed in the end