Optimization of PID parameters with cuckoo search algorithm

This study proposes a novel control method that combines the simplicity and practicality of PID controllers with the strong convergence capabilities of the Cuckoo Search Optimization (CSO) algorithm. Although optimization methods can theoretically guarantee finding global minima and maxima within a defined parameter range, determining the coefficient space for a PID controller is crucial for ensuring system stability, performance and applicability. In the proposed method, the search space for the CSO algorithm is defined using signature formulas. The performance of this innovative approach (σ-CSO-PID) is evaluated through both simulations and real-time implementation on a brushed DC motor. The results demonstrate that PID coefficients optimized using the CSO-PID method within randomly selected ranges outperform those obtained using classical methods in terms of Integral Square Error (ISE), Integral Absolute Error (IAE), and Integral Time Absolute Error (ITAE). With the signature method, the proportional gain is determined by taking into account the system requirements, ensuring system stability. While the CSO-PID method optimized in random ranges showed better performance in simulation studies, it led to excessive overshoot and wind-up conditions in real-time applications. Consequently, the proposed method (σ-CSO-PID) is shown to be superior to classical methods in both simulation and real-time implementation, and more practical compared to the standard CSO-PID approach.