Evaluating Metaheuristic Algorithms and Neural Networks for PID Tuning

This paper aims at comparing various PID tuning methodologies. An armature-controlled DC motor has been modeled in Simulink using its actual parameters for the study. The metaheuristic approaches i.e., Flower Pollination Algorithm and Grey Wolf Optimization have been coded on MATLAB and the tuned parameters have been imported on Simulink to check their response of the system. These algorithms work on ITAE as their objective function. ANFIS controller for tuning PID values is implemented on Simulink. Its training data is imported from Fuzzy Logic Controller's tuned parameters. Finally, the results of the controllers are analyzed and distinguished by values of Rise Time (Tr), Peak Overshoot (Mp), and Peak Time (Tp).