Frequency response masking filter design using an oscillation search genetic algorithm

In this paper, an efficient optimization scheme is proposed for the design of frequency-response masking (FRM) FIR filters in the signed power-of-two (SPoT) space. The scheme is generated by integrating the genetic algorithm (GA) with an oscillation search (OS) algorithm. The optimization process is controlled by the GA. The OS algorithm is developed based on the properties of filter coefficients, which is used as a separate process to refine the elitist individuals. With the help of the proposed approach, all sub-filters can be simultaneously designed in the SPoT space. Simulation results have shown that it is both more efficient and more effective than the GA by requiring less computational cost and identifying higher quality solutions.