ABSTRACT
This paper presents a new methodology for modeling and identifying the dynamics of unknown systems. Assuming that an unknown system belongs to a set of models, the methodology attempts to find the model that matches the characteristics of the unknown system and the best set of parameters for this model. To achieve these goals, a Genetic Algorithm (GA) methodology is utilized as a search strategy due to its efficiency and robustness. The proposed methodology consists of three stages: a model selection stage which is concerned with selecting the subset of models that best matches the input to and the output from the unknown system; a best model selection stage which is concerned with selecting the best model from the pre-selected subset of models from stage 1; and a parameter tuning stage which is concerned with the fine tuning of the best model parameters selected from stage 2. Simulation results are presented which show the effectiveness of the proposed method.
