| Modern industry possesses many characteristic, such as non-linearity, large time lag, time-varying and non-accurate mathematic model and so on. It is difficult to obtain perfect control performance when we adopt conventional PID controller; hence we should find new intelligent control strategy. At present, the combination of fuzzy theory, neural network and genetic algorithm(GA) is gradually showing enormous latent capacity in control field.Genetic algorithm is one of the global search techniques especially useful for the complex optimization problems with a large number of parameters of which an analytical solution is difficult to obtain.Fuzzy Control is an important branch of Intelligent Control.Moreover, it is a kind of control method based on rules, directly adopting language control rules according to the control experiences of local operators or knowledge from experts of this field. But efficiency of Fuzzy Control depends on several key parameters: membership functions, fuzzy control-table and scale factors. Convention methods determining the parameters are man-operated especially based on expert's experience and practical modulation. So subjectivity and randomness exist. Thus some scholars adopt GA to optimize the fuzzy control rules in order to solve the above problem.But we meet a practical problem that we hardly obtain a complete control rules which of all is optimized. In addition, in the final period of GA, the structure of each chromosome is basically uniform. Especially due to a lot of papers regarding step signal as input, this can not ensure optimal control rules cover the entire field well so that it gives rise to loss of control. Although the method ofr 1 o < t <; 40 . . . . , ,using r = \ as input can optimize more control rules5 [0 40process, it can not solve the proposed problem completely.Aiming at the problem, this paper proposed a new method to? ? r , , t^- , f 1 0 < f ï¿¡ 40optimize fuzzy control rules. Firstly, we use r = 1 asF J J [0 40 |
PDF Full Text Request