Non-linear Model Identification Based On Genetic Algorithm

Many systems in modern control project are shown as nonlinear, time-variant and uncertain. To achieve effective control, we should use model identification to obtain the approximate model to the object. Genetic algorithm (GA) is a kind of search algorithms that can be used in model identification and optimality calculation in complex system. It can search the global optimal solution in solution space based on its search mechanism and randomness. It has good robustness and adaptive ability. Basing on having studied immune principle deeply, immunity generation is applied in GA in this paper to improve its optimize. The improved algorithm is applied in identification to a water level control system of a single-capacity water tank so as to obtain the practical characteristic curve of pump. The results validate its superiority of the algorithm based on immunity theory in the use of non-linear model identifying.