Khan, Kashif2023-06-162023-06-162023Khan, K. (2023). Development and Deployment of CAD Integrable Machine Learning Based Modelling Techniques for Advanced RF Devices. School of Engineering and Digital Scienceshttp://nur.nu.edu.kz/handle/123456789/7227This research investigates and advances effective machine learning (ML) methodologies for analyzing the performance of gallium nitride (GaN) high electron mobility transistors (HEMTs). A multilayer perceptron (MLP) framework within an artificial neural network (ANN) is utilized for the behavioral representation of a 2-mm GaN-on-silicon device. It is recognized that ANNs exhibit a dependence on initial weight and bias values. To address this limitation, multiple global optimization algorithms, including grey wolf optimization-particle swarm optimization (GWO-PSO), ant colony optimization (ACO), whale optimization algorithm (WOA), and ant lion optimization (ALO), are integrated into the MLP structure. The models are trained on an extensive dataset encompassing various operating conditions, such as ambient temperatures and bias voltages, within a frequency range of 0.1 to 26 GHz. Subsequently, these models are subjected to temperature interpolation and extrapolation tests, with the aim of determining their level of precision, robustness and stability. A strong correlation between the measured and modeled S-parameters across the entire frequency spectrum attests to the efficacy and resilience of the proposed methodologies. Moreover, the GWO-PSO-assisted ANN outperforms other models, as evidenced by mean squared error (MSE), mean absolute error (MAE), and coefficient of determination (%R2) metrics.enAttribution-NonCommercial-ShareAlike 3.0 United StatesType of access: EmbargoCAD Integrable Machine LearningRF DevicesCapstone Project