Study On Adaptive Characteristic In Fuzzy Control

Intelligent control including artificial intelligence, fuzzy control and neural network is the trend of modern automatic control theory and fuzzy control is the most valuable one among them. In this paper, we do some research on the adaptive characteristic in fuzzy control.The rule base is the core of fuzzy controller. In order to acquire good control performance, it is insufficient to only collect the evident knowlegde of experts. Only we mining the connotative knowledge contained in the input-output data and combine both of them can we construct the suitable rule base. Therefore, it has become a critical problem that how to tune the primary rule base constructed by using the input-output data. In this paper, we firstly discuss some limitations appeared in the course of adjusting the rules using genetic algorithm. Then, we propose our GA-based method which can guarantee the completeness of the rules after tuning and make the membership function of each linguistic variable overlap adaptively. The effectiveness can be seen in two simulations.The solar water heater is a kind of domestic instruments, which is influenced greatly by the change of intensity of the solar radialization. A bottleneck problem during its development is how to make full use of the solar and save electrical energy farthest. In this paper, we first construct the primary rules by lookup table, then tune the rule base adaptively using our GA-based method. As a result, we acquired a fuzzy controller having good control performance which combines solar and electricity intelligently and offers a resolution method for the problem of time-consuming and electric-consuming existed in the electric auxiliary device of the solar water heater. Also, the flow chart of programme designing and system controlling is given.