Parameter Tuning Of PID Controller Based On An Improved Hybrid GA

GA (Genetic algorithm) is an excellent algorithm of optimization. It has taken many researchers' attention since brought foreword in the sixties of 20 century. With much research on it, GA has been used in many fields including function optimization, control systems, machine learning, communication, signal . processing, etc. In recent years, GA was applied to the tuning of PID controller's parameters by some researchers.PID (Proportion, Integral, Differential) control is one of the most wildly used rules in control field because of its simple structure, high reliability and high robustness. To the performance of the controller the tuning of PID parameters is important, but still complex and difficult. The principals and common methods of PID controller parameters tuning were specified in this paper. Compared with the common methods, the tuning method based on GA can get better parameters, with fewer limits. Much research of GA's application to PID parameter tuning has been made, however, how to quicken the speed of the parameter tuning is still a problem unsettled. So this paper brings out a new hybrid GA for PID parameter tuning.In chapter 3, an improved GA based on elite thought (EHGA) is designed. Some new strategies are provided to advance the GA's performance. Then the new algorithm is applied to the optimization of three common functions to test its performance. The analyses of its characteristics and excellence are made based on the results of optimization test. The result shows that EHGA is an efficientalgorithm.In chapter 4, the new algorithm is used in PID controller parameter tuning. The methods of generating the first generation and computing the fitness are introduced. After a simulation with MATLAB, the results is compared with those got by ISTE, SGA and AGA. The simulation shows that the PID controller with its parameters, tuned based on EHGA brought out in this paper has satisfied performance and the tuning speed is far faster than common GA and AGA.In the end, a conclusion is made based on summarizing all the work done by the author. Also. We some problems need to settle in the future is indicated.