Genetic Algorithm-Optimized Convolutional Neural Network Controller for Enhanced Performance of Boost DC-DC Converters

This paper introduces an innovative control strategy for Boost DC-DC converters utilizing a genetic algorithm-optimized convolutional neural network (GA-CNN) to significantly improve dynamic performance and stability. The proposed method incorporates genetic algorithms to identify optimal hyperparameters and network architectures, allowing the CNN controller to efficiently manage voltage regulation and address the nonlinearities commonly associated with Boost converters. By leveraging real-time data on inductor current, capacitor voltage, and output voltage error, the GA-CNN controller exhibits rapid convergence and enhanced transient response when compared to traditional control methods. Comprehensive experimental evaluations and simulations demonstrate the robustness and efficiency of the proposed approach, showcasing its superiority in achieving reduced settling times and minimal overshoot under varying load and input voltage conditions. The findings underscore the potential of the GA-CNN methodology for high-performance applications in power electronics, paving the way for further advancements in adaptive control strategies within this domain.