CONSTRUCTING THE FUNCTIONS FOR SEMANTIZATION AND DESEMANTIZATION IN HEDGE-ALGEBRAS-BASED REASONING METHODS

Abstract

To design the controller according to hedge algebra to effective and accurate working, it is necessary for turning the parameters so that the model truly reflects the system's rules. In this paper, the author proposes to use the nonlinear semantization and desemantization nonlinear functions instead of the linear normalization and desemantization nonlinear functions. Nonlinear functions are proposed for both semantization and desemantization that are exponential with the same exponential parameter. This proposal was applied to design and simulate a controller for a furnace object. In addition, the genetic algorithms is also used to optimize these exponential parameters. Comparing the simulation results with the controllers, we found that when using the semantization and desemantization nonlinear function gave a better control response when using the linear normalization functions.