Optimizing Flag-Shaped Patch Microstrip Antenna Performance Using Tree-Based Learning Techniques

Microstrip patch antennas are essential in modern communication systems due to their compact size, low profile, and ease of fabrication. However, optimizing their physical parameters to achieve a target resonant frequency often requires computationally expensive simulations. Machine learning (ML) provides an effective alternative by predicting antenna performance based on structural parameters. This study investigates Random Forest (RF) and Extreme Gradient Boosting (XGBoost) models, both with and without Genetic Algorithm (GA) optimization, for estimating the resonant frequency of a flag-shaped microstrip patch antenna. The dataset was constructed by systematically varying patch and slot dimensions and substrate characteristics. GA was utilized to optimize model hyperparameters and improve accuracy. Models were evaluated using Root Mean Square Error (RMSE) and R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>, revealing that RF-GA achieved the best performance, followed by XGBoost-GA. The study also analyzes the computational trade-offs introduced by GA. These findings demonstrate that ML-based predictive modelling, particularly when combined with optimization algorithms, can significantly streamline the antenna design process. This work is relevant to applications in antennas, microwave engineering, and emerging technologies such as 5G and AI-driven communication systems.