A study on generation of fuzzy control rules based on FFOTS

Fuzzy control rules are the very core of a fuzzy controller. Most traditional fuzzy controllers consider the current status of systems only and the generation of them is written by a human expert, which accounts for its slow adjustment and large shocks around stable point. This paper, from the angle of combination of system historical status trend and the current status of the system, proposes a new kind of fuzzy control rules and gives a specific method of its generation from industrial data by using data mining techniques. At last it is proved by MATLAB simulation experiments that this new kind of fuzzy controller reduces the system´s adjustment time and overshoot by a large extent compared with the traditional fuzzy controller.