Evolutionary Computation And Its Applications In Adaptive Filtering

Evolutionary Computation is an important branch of artificial intelligence research that has been given broad attention in recent years; it is also a main part of intelligent information processing as well. As an optimization algorithm based on the theory of biologic evolution, the most outstanding advantage of Evolutionary Computation is its strong global optimizing capability as compared with other optimization algorithms. In the last decade, more and more people pay attention to the methods of the intelligence computation to solve the problem of the engineering optimization. In signal processing field, there has been growing interest in optimization algorithms of the adaptive filter. LMS algorithm is commonly used in the optimum design of the adaptive filter, since LMS adaptive algorithm is a simple algorithm and can be realized easily, but the convergence behavior and maladjustment of the LMS algorithm is seriously affected by the step-size, and the optimum parameter of step-size can not be calculated easily. Aimed at solving this problem, an improved fast Evolutionary Programming is proposed in this thesis based on the research of the Evolution Computation. Adaptive FIR filter as an example, In the design of adaptive filter, the improved fast Evolutionary Programming is used; in this algorithm, the individual is checked when it is generated, if the individual is illegal, then, give it up; On the other hand, Cauchy mutation takes the place of Gaussian mutation to speed up of the convergence, and it can reach the global optimization and converge quickly; This algorithm is not depended on any parameters and function, it only needs the error values. As a result, it has some advantages as compared with the LMS algorithm, thus is expected being useful in engineering applications. We can get a good result by the simulation and indicate the validity of the algorithm by many experiments.