Identification of fuzzy systems by means of an auto-tuning algorithm and its application to nonlinear systems

The study concerns a design procedure of rule-based systems. The proposed r ule-based fuzzy modeling implements system structure and parameter identifi cation in the efficient form of "IF..., THEN..." statements, and exploits t he theory of system optimization and fuzzy implication rules. Two types of methods for rule-based fuzzy modeling are studied. This classification conc erns the form of the conclusion part of the rules that can be either consta nt or formed by some linear functions. Both triangular and Gaussian-like me mbership function are studied. The optimization hinges on an auto-tuning al gorithm that covers a modified constrained optimization method known as a c omplex method. The study introduces a weighted performance index (objective function) that helps achieve a sound balance between the quality of result s produced for the training and testing set. This methodology sheds light o n the role and impact of different parameters of the model on its performan ce tin-particular, the mapping and predicting capabilities of the rule-base d computing). The study is illustrated with the aid of several representati ve numerical examples. (C) 2000 Elsevier Science B.V. All rights reserved.