An Adaptive Frequency Sweep Algorithm for Broadband EM Simulation Incorporating Random Matrix Transformation in the Loewner Framework

Abstract

In this work, we developed an adaptive frequency sweep algorithm exploiting the Loewner matrix model. This algorithm requires two Loewner matrix models. In addition to the conventional Loewner matrix model, we create a second model by employing a random matrix transformation on the Loewner matrix. The discrepancy between these two models is assessed, and the frequency point associated with the highest discrepancy is designated as the subsequent simulation point. To confirm the accuracy and speed of this algorithm, two numerical examples are performed using MATLAB R2025a. The proposed algorithm and the MATLAB R2025a frequency sweep object have been compared in this work. The results show that the proposed Loewner matrix-based adaptive frequency sweep method works incredibly well in terms of speed and accuracy.