The Prominent Stochastic Optimization Algorithm Settings for the Self-Regulating Feedback Control

This research demonstrates a direct way to obtain determinant variables of the optimization assessment, enabling the optimal control tuning for the closed-loop process. Stochastic optimization approaches such as Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) is analyzed and compared with the performance of manually calculated Proportional-Integral-Derivative (PID) tuning, which are calculated from the defined algorithms. Nevertheless, the performance of both optimizations might get restricted without proper settings for its determinant variables. This research examines the determinant variables for the optimization assessment. The Upper and Lower boundaries (UB, LB), mutation rate (mu), damping ratio (wdamp), maximum iteration (MaxIt) and population size (nPop) are analyzed and discussed in detail. In the validation, a Level Control (SE-207) module produced the curve responses by applying the controller settings of all the manually calculated PID, GA and PSO algorithms. The result implies improvements in GA and PSO as compared with the manually calculated PID tunings. Moreover, PSO triggered higher overshoots than the GA even though the error values from both optimizations were extremely close. Whereas, GA offers more robust and stabilized responses therefore is favorably selected for the operation of the aforementioned physical module.