Design of FIR Filter Based on Adaptive Gridding Partition Ant Colony Optimization

To approximate the response of the ideal filter with a finite impulse response sequence is the purpose of designing FIR filter. In frequency sampling FIR filter designing, finding the optimal sampling value in transitional-band can effectively reduce the approximation errors. Sampling value found by Genetic Algorithms(GA) reduce the errors in a certain degree. But for the limit of binary coding, the precision of sampling value could be improved by solving in continuous domain. In this paper, we suggest a novel method Adaptive Gridding Partition Ant Colony Optimization(AGPACO) to find the optimal value in transition-band. The frequency sampling FIR filter designed by AGPACO and GA are simulated respectively. The result shows FIR filters designed by AGPACO have higher minimum stop-band attenuation than FIR filters designed by GA.