A Unified Automated Parametric Modeling Algorithm Using Knowledge-Based Neural Network and ${l}_{1}$ Optimization

Knowledge-based neural network modeling techniques using space-mapping concept have been demonstrated in the existing literature as efficient methods to overcome the accuracy limitations of empirical/equivalent circuit models when matching new electromagnetic data. For different modeling problems, the mapping structures can be different. In this paper, we propose a unified automated model generation algorithm that uses <inline-formula> <tex-math notation="LaTeX">$\boldsymbol {l}_{\mathbf {1}}$ </tex-math></inline-formula> optimization to automatically determine the type and the topology of the mapping structure in a knowledge-based neural network model. By encompassing various types of mappings of the knowledge-based neural network model in the existing literature, we present a new unified model structure and derive new sensitivity formulas for the training of the unified model. The proposed <inline-formula> <tex-math notation="LaTeX">$\boldsymbol {l}_{\mathbf {1}}$ </tex-math></inline-formula> formulation of modeling can force some weights of the mapping neural networks to zeros while leaving other weights as nonzeros. We utilize this feature to allow <inline-formula> <tex-math notation="LaTeX">$\boldsymbol {l}_{\mathbf {1}}$ </tex-math></inline-formula> optimization to automatically determine which mapping is necessary and which mapping is unnecessary. Using the proposed <inline-formula> <tex-math notation="LaTeX">$\boldsymbol {l}_{\mathbf {1}}$ </tex-math></inline-formula> optimization method, the mapping structure can be determined to address different needs of different modeling problems. The structure of the final knowledge-based model can be flexible combinations of some or all of linear mapping, nonlinear mapping, input mapping, frequency mapping, and output mapping. In this way, the proposed algorithm is more systematic and can further speed up the knowledge-based modeling process than existing knowledge-based modeling algorithms. The proposed method is illustrated by three microwave filter modeling examples.

[1]  Weicong Na,et al.  Unified automated knowledge-based neural network modeling for microwave devices , 2015, 2015 IEEE MTT-S International Conference on Numerical Electromagnetic and Multiphysics Modeling and Optimization (NEMO).

[2]  E. Hammerstad,et al.  Accurate Models for Microstrip Computer-Aided Design , 1980, 1980 IEEE MTT-S International Microwave symposium Digest.

[3]  Qi-Jun Zhang,et al.  Evolutionary Neuro-Space Mapping Technique for Modeling of Nonlinear Microwave Devices , 2011, IEEE Transactions on Microwave Theory and Techniques.

[4]  Chao Zhang,et al.  Parametric Modeling of EM Behavior of Microwave Components Using Combined Neural Networks and Pole-Residue-Based Transfer Functions , 2016, IEEE Transactions on Microwave Theory and Techniques.

[5]  M.L. Strydom,et al.  Reliability and efficiency in computational electromagnetics: How CEM software meets the future needs of engineers , 2007, 2007 IEEE Antennas and Propagation Society International Symposium.

[6]  Kai Chang,et al.  Microstrip Elliptic-Function Low-Pass Filters Using Distributed Elements or Slotted Ground Structure , 2006, IEEE Transactions on Microwave Theory and Techniques.

[7]  F.M. Ghannouchi,et al.  Dynamic behavioral modeling of 3G power amplifiers using real-valued time-delay neural networks , 2004, IEEE Transactions on Microwave Theory and Techniques.

[8]  John W. Bandler,et al.  Microwave device modeling using efficient l1 optimization: a novel approach , 1986 .

[9]  Qi-Jun Zhang,et al.  Parametric Modeling of Microwave Passive Components Using Sensitivity-Analysis-Based Adjoint Neural-Network Technique , 2013, IEEE Transactions on Microwave Theory and Techniques.

[10]  V. Gutierrez-Ayala,et al.  EM-Based Monte Carlo Analysis and Yield Prediction of Microwave Circuits Using Linear-Input Neural-Output Space Mapping , 2006, IEEE Transactions on Microwave Theory and Techniques.

[11]  M.C.E. Yagoub,et al.  Efficient analytical formulation and sensitivity analysis of neuro-space mapping for nonlinear microwave device modeling , 2005, IEEE Transactions on Microwave Theory and Techniques.

[12]  D. Depriest,et al.  Understanding Industrial Designed Experiments , 1993 .

[13]  Kaj Madsen,et al.  Combined LP and Quasi-Newton Methods for Nonlinear $l_1 $ Optimization , 1985 .

[14]  J.E. Rayas-Sanchez,et al.  EM-based optimization of microwave circuits using artificial neural networks: the state-of-the-art , 2003, IEEE Transactions on Microwave Theory and Techniques.

[15]  John W. Bandler,et al.  A nonlinear l_1 optimization algorithm for design, modeling, and diagnosis of networks , 1987 .

[16]  José E. Rayas-Sánchez,et al.  Neural input space mapping optimization based on nonlinear two-layer perceptrons with optimized nonlinearity , 2010 .

[17]  J.W. Bandler,et al.  Neuromodeling of microwave circuits exploiting space mapping technology , 1999, 1999 IEEE MTT-S International Microwave Symposium Digest (Cat. No.99CH36282).

[18]  Jose E. Rayas-Sanchez,et al.  Neural input space mapping optimization based on nonlinear two-layer perceptrons with optimized nonlinearity: Neural Input SM Opt. with Nonlinear 2LPs , 2010 .

[19]  Qi-Jun Zhang,et al.  Neural Networks for RF and Microwave Design , 2000 .

[20]  Fang Wang,et al.  Knowledge based neural models for microwave design , 1997, 1997 IEEE MTT-S International Microwave Symposium Digest.

[21]  G.A.E. Vandenbosch,et al.  Quality Assessment of Computational Techniques and Software Tools for Planar-Antenna Analysis , 2009, IEEE Antennas and Propagation Magazine.

[22]  D. Wisell,et al.  Wide-band dynamic modeling of power amplifiers using radial-basis function neural networks , 2005, IEEE Transactions on Microwave Theory and Techniques.

[23]  Sushanta K. Mandal,et al.  ANN- and PSO-Based Synthesis of On-Chip Spiral Inductors for RF ICs , 2008, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems.

[24]  J.W. Bandler,et al.  Space Mapping With Adaptive Response Correction for Microwave Design Optimization , 2009, IEEE Transactions on Microwave Theory and Techniques.

[25]  S. Koziel,et al.  A Space-Mapping Framework for Engineering Optimization—Theory and Implementation , 2006, IEEE Transactions on Microwave Theory and Techniques.

[26]  S. Koziel,et al.  Space mapping , 2008, IEEE Microwave Magazine.

[27]  Qi-Jun Zhang,et al.  Simple and Efficient High-Dimensional Parametric Modeling for Microwave Cavity Filters Using Modular Neural Network , 2011, IEEE Microwave and Wireless Components Letters.

[28]  Geoffrey E. Hinton,et al.  Learning internal representations by error propagation , 1986 .

[29]  J.W. Bandler,et al.  Space mapping: the state of the art , 2004, IEEE Transactions on Microwave Theory and Techniques.

[30]  John W. Bandler,et al.  Neural space-mapping optimization for EM-based design , 2000 .

[31]  R. Jansen,et al.  Accurate Wide-Range Design Equations for the Frequency-Dependent Characteristic of Parallel Coupled Microstrip Lines , 1984 .

[32]  Qi-Jun Zhang,et al.  Smart Modeling of Microwave Devices , 2010, IEEE Microwave Magazine.

[33]  D.E. Root,et al.  A behavioral modeling approach to nonlinear model-order reduction for RF/microwave ICs and systems , 2004, IEEE Transactions on Microwave Theory and Techniques.

[34]  R. Esfandiari,et al.  Design of Interdigitated Capacitors and Their Application to Gallium Arsenide Monolithic Filters , 1983 .

[35]  Weicong Na,et al.  A review of neural network based techniques for nonlinear microwave device modeling , 2016, 2016 IEEE MTT-S International Conference on Numerical Electromagnetic and Multiphysics Modeling and Optimization (NEMO).

[36]  Feng Feng,et al.  Parallel Space-Mapping Approach to EM Optimization , 2014, IEEE Transactions on Microwave Theory and Techniques.

[37]  John W. Bandler,et al.  Microwave Device Modeling Using Efficient l/sub 1/Optimization: A Novel Approach , 1986 .

[38]  Qi Jun Zhang,et al.  Automated Knowledge-Based Neural Network Modeling for Microwave Applications , 2014, IEEE Microwave and Wireless Components Letters.

[39]  M.C.E. Yagoub,et al.  A robust algorithm for automatic development of neural network models for microwave applications , 2001, 2001 IEEE MTT-S International Microwave Sympsoium Digest (Cat. No.01CH37157).