Parametric modeling of electromagnetic (EM) behavior has become important for EM design optimization of microwave components. This paper provides an overview of the advance in parametric modeling using combined neural networks and pole-residue-based transfer functions. EM sensitivity information is further utilized to speedup model development by reducing the number of training data required for developing the model. A microwave component example is used to illustrate this advanced techniques.

Additional Metadata
Keywords Electromagnetic modeling, Neural networks, Pole-residue-based transfer functions, Sensitivity
Persistent URL dx.doi.org/10.1109/NEMO.2018.8503497
Conference 2018 IEEE MTT-S International Conference on Numerical Electromagnetic and Multiphysics Modeling and Optimization, NEMO 2018
Citation
Feng, F. (Feng), Zhang, C. (Chao), Na, W. (Weicong), Zhang, J. (Jianan), Zhang, W. (Wei), Zhao, Z. (Zhihao), … Zhang, Q.J. (2018). Recent Advances in Parametric Modeling Using Pole-Residue-Based Neuro-Transfer Functions. In 2018 IEEE MTT-S International Conference on Numerical Electromagnetic and Multiphysics Modeling and Optimization, NEMO 2018. doi:10.1109/NEMO.2018.8503497