Tuning evolvable PID controllers through a clonal selection algorithm

Abstract

This work discusses the use of an evolvable proportional-integral-derivative (PID) controller that consists of a PID controller hardware whose gains can be set, for example, by evolutionary computation techniques, such as genetic algorithms. Due to PID controllers' widespread use in industry, tuning procedures for them are always a topic of interest. An evolutionary immune inspired algorithm, named clonal selection algorithm, is used for tuning the controller so that closed-loop step response specifications are satisfied. By using this procedure, designers need only specify the desired closed-loop response. Experiments with different processes indicate that the gains obtained through clonal selection algorithms may provide better responses than those obtained by the classical Ziegler-Nichols method. Moreover, the clonal algorithm is capable of generating adequate gains for systems where classical rules are not applicable.