PID Optimization Method Based on Heuristics Algorithm And Q-Learning

Abstract

Proportional-integral-derivative (PID) control is a feedback control algorithm that adjusts the output signal through a combination of three control terms, proportional, integral, and differential, and is used to stabilize the actual output of the system close to the desired value. Heuristic algorithms are a class of intelligent problem-solving methods that optimize the solution using rules and trial-and-error learning; Q-learning is a reinforcement learning algorithm that focuses on decision-making in the Markovian decision-making process and achieves the optimal strategy by continuously learning and updating the value function. As people's requirements for the accuracy of various control systems increase, the requirements for the stability and robustness of PID control algorithms also increase. The traditional Ziegler-Nichols method to find PID parameters can no longer meet people's needs well, and people want to find more stable PID parameters faster. This article introduces different heuristic algorithms, such as particle swarm optimization algorithm and Q-learning algorithm for optimizing PID parameters in different systems. The effect of optimization of different algorithms is also analyzed.