Parameter extraction and optimization using Levenberg-Marquardt algorithm

Abstract

Parameter extraction is an important part of model development. The goal of parameter extraction and optimization is to determine such values of device model parameters that minimize the differences between a set of measured characteristics and results obtained by evaluations of the device model. This minimization process is often called fitting of model characteristics to the measurement data. The objective of this paper is presenting an extraction and optimization method, using Levenberg-Marquardt algorithm, for a set of electrical parameters of the EKV MOSFET Model. The Levenberg-Marquardt algorithm is an efficient and popular damped least square technique. This algorithm is a combination between the steepest gradient descent and the Gauss-Newton algorithms. All implementations are carried out on Matlab environment.