PID controller tuning using hybrid optimisation technique based on Box's evolutionary optimisation and teacher-learner-based-optimisation

Abstract

In this paper, a hybrid optimisation technique based on Box's evolutionary optimisation and teacher-learner-based-optimisation (BEO-TLBO) is proposed for proportional-integral-derivative (PID) controller tuning of level control of three-tank system. The integral-square-error (ISE) of unit step response is minimised for obtaining optimal controller parameters. The ISE is designed in terms of alpha and beta parameters. In BEO-TLBO, a global search is first carried out over the entire search space to determine the set of desired controller parameters using teacher-learner-based-optimisation (TLBO). The search is then refined in the second stage using Box's evolutionary optimisation (BEO). The results obtained using BEO-TLBO are compared with other existing techniques. Computer simulations reveal that the hybrid optimisation based approach meets the desired specifications with greater accuracy as compared to the other existing methods.