A novel evolutionary programming for adaptive filter based on learning mutation

The convergence algorithm is always the research hotspot in the optimum design of the adaptive filter. In this paper, a novel evolutionary programming is introduce for the design of the adaptive filter, in the algorithm, a new learning mutation operator is put forward, it have a fast mutation capability, which can improve premature convergence of traditional evolutionary programming, and it can overcome the convergence behavior error and maladjustment of the LMS algorithm, on the other hand, this algorithm is not depended on any parameter. After many experiments, we can get a good result by the simulation curve, which indicate the validity of the novel algorithm.