An improved data-complementing method via fuzzy rough sets for fuzzy-relationship matrix modeling and applications

Abstract

An improved data-complementing algorithm using fuzzy rough sets is presented in this work. The fuzzy systems with incomplete data and similarity matrices are defined for increasing accuracy of a fuzzy relationship matrix. A complete sampled-data system is formulated by complementing the controller's input and output information. Then, a fuzzy relationship matrix based on a semi-tensor product is established. This method is applied to air-conditioning control systems for an indoor thermal environment. A complete fuzzy-relationship matrix model for the fuzzy controller is built after the experimental data has been complemented. Compared with the model established using the incomplete data, simulation studies show that the fuzzy controller established using complete data can greatly improve the control accuracy of the indoor comfortability.