Step Size Optimization of LMS Algorithm Using Genetic Algorithm in System Identification

System identification is the art and science of building mathematical models of dynamic systems from observed input-output data. This paper combines Genetic algorithm and LMS algorithm to describe the application of a Genetic Algorithm (GA) to the problem of parameter optimization for an adaptive Finite Impulse Response (FIR) filter. LMS algorithm computes the filter coefficients and GA search the optimal step-size adaptively. Because step-size influences on the stability and performance, so it is necessary to apply method that can control it.. However, the statistical Least Mean Squares method is faster than the genetic algorithm. For this reason we suggest using the genetic algorithm for off-line applications, and the statistical method for on-line adaptation. A hybrid method combining the advantages of both methods is proposed for real world applications. In Genetic algorithm, we have used Roulette wheel Selection, Arithmetic Crossover, Uniform Mutation& .the simulation results of the GA were compared to the traditional fixed step size LMS algorithm.