Exploring optimal controller parameters for complex industrial systems

Tuning controller parameters to achieve desired system performance is challenging, especially for complex systems. Many heuristic methods are proposed to solve the problem. Because there are many system performance indices, such as response time and overshoot, it is difficult for these methods to achieve desired system performance due to system complexity, noise and uncertainties etc. This paper proposes an automated parameter tuning method, Gaussian Process Regression surrogated Bayesian Optimization Algorithm (GPRBOA), based on the required system performance for complex industrial systems. Because proportional-integral-derivative (PID) controller is widely used in industry, it is used as an example to demonstrate how the proposed method works. GPRBOA is applied to optimize the PID parameters by iteratively updating the system model and optimizing the system performance. Simulations have been performed and the results demonstrate the effectiveness of the proposed method.