PID Optimization Of Uncertain Object Based On Genetic Algorithm

In the practical applications of control engineering, PID controller is widely usedfor easy operation and easy calculation. However, there are many uncertainties in thereality industrial production of the complex system. The control system even will besome mutations phenomenon. When the system has many uncertainties, the traditionalPID parameters setting method may be unable to meet the design requirements.However, strictly speaking, there are no control system do not have uncertainties in thepractical environment. So this paper studies the optimization problem of uncertainobjects whose model parameters are interval number of PID controller. The maincontent of this study are as follows:(1) This paper briefly introduces the principle of PID controller and thedevelopment status of genetic algorithms. It shows the necessity of academic study forthe PID optimization of uncertain object.(2) This paper analyzes a class of uncertain optimization problems for whosemodel parameters is the interval number. First this paper analyzes the mathematicalexpression of these propositions, then it approaches solving ideas of the optimizationproblem from the mathematical theory, and finally converts to the problem raisedMINIMAX optimization problem.(3) In order to better solve own existence flaw of the MINIMAX optimizationproblems, it’s proposed an improved hybrid genetic algorithm which making theproposition can accurately get the global optimal solution. This new algorithmcombines the improved genetic algorithm and the simplex optimization algorithm.Improved genetic algorithm uses real-coded. Then it adopts elitist strategy the "elitist"strategy, and it follows by the implementation of genetic manipulation in accordancewith the modified roulette wheel selection operator, arithmetic crossover and improvedadaptive Gaussian mutation operator. It provides a guideline to the uncertain object PIDoptimization solution.(4) In the control engineering, uncertainty parameter is very widespread. For thecase of uncertainty parameter, the paper first combed the integrated time and absoluteerror, control volume, rise time, overshoot as the optimal target. Then it’s proposed touse the previously mentioned MINIMAX optimization algorithm optimizing the PID parameters of uncertain objects. Finally it gets satisfactory optimization results.