Improved Genetic Algorithm And Its Application In PID Controller Parameter Optimization

PID control is the most practical application of the most widely used industrial process control strategy, while their study has yielded fruitful results. In the industrial process, more than 95%of the controller is PID controller. Especially since 1993, due to the development of modern control theory divorced from the actual engineering and modern industrial process control system for the high demands made on the PID Control of the outbreak was the trend. PID controller parameters optimization and performance tuning is critical for the control system. At present, domestic and foreign scholars in the nominal model of the object under the control of PID controller tuning and optimizing the design, but in actual control system, due to the work environment, aging components, the structure of the geometric error, materials and measurements factors, the control object parameter uncertainty is an objective reality. Control object-based model under nominal controller tuning and optimization, object parameter in the control of large changes, it is difficult to meet the control requirements. This article parameters for the control of the uncertainty prevailing the field of a class of linear PID control system, combined with the characteristics of randomized algorithms, an improved genetic algorithm and its application to the vehicle by-wire system, the design of PID controller optimization, mainly to do the following work:First, the use of Hermite-Biehler theorem and the generalized Hermite-Biehler theorem of genetic algorithms to determine the PID parameters optimization space. In the general framework of the traditional genetic algorithm, the introduction of random algorithm, traditional genetic algorithm improved the objective function, the objective function is no longer a time-domain, frequency domain, integral error, or their combination, determine the function, but the random function the plant model parameters change, to meet the target probability function. Finally the improved genetic algorithm for a class of parameter uncertainty PID linear control system with PID controller parameter tuning of the numerical simulation results compared with the traditional genetic algorithms, large changes in the parameters of the controller would still be better dynamic performance quality, validated algorithm.Secondly, the vehicle steering system control object modeling, application of improved genetic algorithm to optimize parameters of the PID control and tuning, simulation results compared with the traditional genetic algorithm, large changes in the parameters of the controller, the dynamic performance can still get better quality, and highly robust.Finally, a summary of the text, combined with genetic algorithm in terms of my study and research work on the next prospect.