Optimizing Flag-Shaped Patch Microstrip Antenna Performance with Machine Learning Models

Microstrip patch antennas are vital in modern communication due to their compact, lightweight design and easy fabrication. However, optimizing their parameters for a target resonant frequency requires time-intensive simulations. Machine learning (ML) offers an efficient alternative by predicting antenna performance from structural parameters. This study evaluates Support Vector Machine (SVM) and Random Forest (RF) models, with and without Genetic Algorithm (GA) optimization, for estimating the resonant frequency of a flag-shaped microstrip patch antenna. The dataset was generated by systematically varying patch and slot dimensions and substrate properties. GA optimized model hyperparameters to enhance accuracy. Models were assessed 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 GA significantly improves prediction accuracy, with RF-GA performing best. The study also examines GA's computational trade-offs. These findings highlight ML-driven predictive modelling as a faster, reliable alternative to traditional simulation-based antenna design.