The Genetic Fuzzy Control And Applied Research In Optimal Control

The selection and optimization of the fuzzy controller’s membership functions andfuzzy control rules are lack of self-learning ability and knowledge acquisition means, andthe genetic algorithm has the characteristics of adaptive, heuristic, probabilistic and iterativeglobal convergence, so the application of good search capability of genetic algorithm tooptimize the fuzzy controller, can achieve good control effect. This paper uses the improvedgenetic algorithm to optimize fuzzy controller parameters.Firstly, this article analyzes and summarizes to the traditional genetic algorithm, andobtains the basic structure and characteristics of the basic genetic algorithm, based on this,advances an improved genetic algorithm (IGA): Within a population into decimal encodingstrategies and classification and the choice of replication operator to expand the diversity ofpopulation and individual genes contain the amount of information; Split operator isintroduced to avoid the genetic algorithm in the optimization process into a local optimalsolution, at the same time to the crossover operator and mutation operator to do thecorresponding adjustment and improvement. By choosing two typical mathematicalfunctions to test for the improved algorithm, the experimental results show that: comparedwith the basic genetic algorithm, the improved algorithm on the accuracy and success rateof searching the optimal solution has certain superiority. And put into use the improvedalgorithm to optimize the membership functions of fuzzy controller, scale factor andquantification factors and control rules to realize comprehensive optimization of fuzzycontroller. The MATLAB simulation results show that, compared with the basic geneticalgorithm, this algorithm has better ability of evolutionary optimization, and the optimizedfuzzy controller has a good control performance.On the basis of the basic genetic algorithm, this paper concludes that the basicstructure and characteristics of the general double population genetic algorithm, andimproves the double population genetic algorithm, between the populations, migrationstrategy is used to speed up the convergence rate of the individual, and enhance thesearching capability of the algorithm, avoiding algorithms fall into the predicament of theprecocious. In addition, this article in view of the wide ranging triangle membershipfunction of fuzzy controller can get better control effect, of this problem, the compressionfactor is used into the membership functions, this method can greatly reduce the code sizeand increase the speed of search for parameter optimization. The improved algorithm usesto optimize membership function and fuzzy rules, scale factor and quantification factors, realized the overall optimization of fuzzy control system. As can be seen from theexperiments results, the improved double population genetic algorithm to optimize thefuzzy controller has smaller overshoot and better response characteristics, the system hasgood anti-interference characteristics, strong robustness, better adaptive ability, and hasobtained the satisfactory control effect. Finally, merge the improved algorithm of the fuzzycontroller is used for single-stage inverted pendulum control system simulation and realcontrol system, Experimental results show that the controller can stabilization control of theinverted pendulum system, has the good real-time control ability.