An Improved Clonal Selection Algorithm Based Optimization Method for Iterative Learning Control Systems

In this paper an improved Clonal Selection Algorithm (CSA) is proposed as a method to implement optimality based Iterative Learning Control algorithms. The strength of the proposed method is that it not only can cope with non-minimum phase plants and nonlinear plants but also can deal with constraints on input conveniently by a specially designed mutation operator. In addiction, because more priori information was used to decrease the size of the search space, the probability of the clonal selection algorithm converging rapidly to a global optimum was increased considerably. Simulations show that the convergence speed is satisfactory regardless of the nature of the plant.