Abstract
A modified fuzzy relational modeling structure is presented in this paper by proposing a fuzzy relational inner composition. This leads to an integrated framework for representing multilevel fuzzy relational equations which is able to provide multi-dimensional models of complex systems very effectively. A potentially online derivative-based identification algorithm is also presented for the proposed model. The identification procedure is based on the special t-norm-based and t-conorm-based operations leading to an easy and fast implementation of the proposed algorithm. The proposed modeling structure is customized and used for two different applications. The simulation results show high modeling capability of the proposed fuzzy relational model.