Genetic Algorithm for the Optimization of an Electrolyte Flow Controller in a Non-conventional Manufacturing Method

This work describes the optimization of a fuzzy controller using a genetic algorithm based on Darwin's principle of natural selection and Mendel's inheritance laws. An elitist selection, blend alpha crossover, and uniform mutation are implemented, with crossover rates of 60 and 80 %, evaluating its performance statistically. The chromosome with the best fitness is implemented in an electrolyte flow controller for an electrochemical machining process. The result is compared with a classical Proportional-Integral-Derivative controller tuned with Ziegler-Nichols, showing that the optimized regulator achieves better performance, with a 7.95 % overshoot, a steady-state error of 0.1151, and a settling time of 4.8 s