GA-Fuzzy Identification of Chaotic Systems
Junhua Liu · Acta Simulata Systematica Sinica · 2004
This paper presents a fuzzy logic-based approach optimized by genetic algorithms (GA) to identify the unknown or uncertain chaotic dynamic systems. Only output data of the underlying system are required to establish the Takagi-Sugeno-Kang (TSK) fuzzy model. The suggested algorithm is composed of two steps. First fuzzy clustering is applied to obtain a compact initial TSK fuzzy model from the data for training. Then this model is optimized by a real-coded GA subjected to constraints that maintain the semantic properties of the rules. A qualitative evaluation of the identified models is made with the reconstruction of an attractor by the identified models, and a quantitative evaluation of the identified models is made with calculation of the Lyapunov exponents of the identified models, too. Simulations show that the identified models can approach the original chaotic dynamic systems and extract the dynamical characteristics of the original chaotic systems.