Fuzzy graphic rule network and its application on water bath temperature control system

In this paper, a novel fuzzy neural network called fuzzy graphic rule network (FGRN) is presented. FGRN has a simple structure and the initial value of its parameters can be easily chosen based on human experience. These parameters are then adjusted during system operation using steepest descent technique. The step length or learning rate is adaptively selected to ensure system stability. As an example, here we employ FGRN as a controller for controlling the temperature of the water bath. Even though the plant´s characteristic is highly nonlinear, it is found from the simulation that the FGRN controller can give satisfactory results.