GaN Power Transistor Modeling Using Global Optimization Based Artificial Neural Networks

Abstract

Recently, Global Optimization algorithms are becoming an efficient tool to solve complicated problems for various applications. Modern optimization algorithms like Black Hole Optimization (BHO) and Social Spider Optimization (SSO) algorithms are widely used to solve different problems. This paper is demonstrating the applicability of these techniques for training an artificial neural network (ANN) and finding optimal values for the weights and biases. The proposed optimization-based ANN model has been utilized to model the IV characteristics of Gallium Nitride High Electron Mobility Transistor (GaN HEMT). A very good fitting is obtained with measurements which shows the validity of both BHO and SSO for this application.