PID Controller Tuning For An Inverted Pendulum Through Meta-Heuristic Algorithms: Firefly And Simulated Annealin

Abstract

In industrial plants a large percentage of the controllers is based on the PID Control Algorithm, due to its simplicity and robustness, it is known that the performance of the Plant may be affected in quantity and quality of the product due to inadequate determination of the controller parameters, even harming the actuators themselves. In this work, an alternative to classical methods of tuning of these parameters is shown: the metaheuristic algorithms, which belong to the branch of Informed Search within Artificial Intelligence, whose objective is to optimize a certain cost function avoiding maximums or local minimums. We have considered the linearized model (for practical purposes) of an inverted pendulum system with a sliding carriage for the application of two algorithms: simulated annealing and fireflies; which had as a higher rank the one determined by the Routh-Hurwitz Criterion for all the specifications of time domain performance. In the results it was possible to determine that the optimized parameters of the PID controller did not have a significant difference and the speed of convergence was fast, this would allow us to conclude that it is an additional option for the tuning of control loops.