Computing optimum coefficients of PID controller with genetic algorithm

Abstract

This paper presents a new method with genetic algorithm (GA) to tune coefficients of PID controller base on minimizing a cost function. Often, methods that exist to find coefficients of PID controllers are based on experience and analytical methods. By development of microprocessors and increasing their speed, if we use a powerful and fast algorithms, numerical optimization has a further advantage than analytic methods. To find PID coefficients, time has an essential role; therefore use of fast algorithms is very important. By using GA, writing the formulas are not necessary and only determination of essential subject in design is sufficient. Therefore GA could be work with fast speed to solve optimization problems. To test ability and performance of GA in finding PID controller coefficients, we used it for a third order system as an practical process example and achieved acceptable results in our implementation with GA. Integral Absolute Error (IAE) criterion used to approve effectiveness of GA in comparison with Integrated Square Error (ISE) method and we found approximately 11.11% error rate improvement than ISE.