Nonlinear dynamical system identification based on evolutionary interval type-2 TSK fuzzy systems

For an interval type-2 fuzzy logic system (IT2FLS), its structure and parameters are learned simultaneously by using evolutionary strategy in this paper. Then gradient descent (GD) and recursive least-squares with forgetting factor (FFRLS) algorithms are employed to optimize the parameters of the IT2FLS. Furthermore, a more efficient type-reduction method, called enhanced iterative algorithm with stop condition (EIASC), is utilized. Finally, an evolutionary interval type-2 TSK fuzzy logic system (EIT2FLS) is developed. The results of applying EIT2FLS to nonlinear systems identification problems demonstrated the superiority of the developed EIT2FLS to existing methods.