Abstract:A neural space mapping optimization algorithm based on nonlinear two layer perceptrons (2LP) is described in this article. This work is an improved version of the Neural Space-Mapping (NSM) algorithm that uses three layer perceptrons (3LP) to implement a nonlinear input mapping function at each iteration. The new version uses a nonlinear 2LP whose nonlinearity is automatically regulated with classical optimization algorithms. Additionally, the new algorithm uses a different optimization method to train the SM-… Show more
“…This technique uses neural networks to map the voltage or current signals of the existing device model into that of the device data. The Neuro-SM method can be applied to not only the simple DC and S-parameters modeling of nonlinear devices, but also the complex large-signal modeling [6,7,8]. In [9], a dynamic neural network is used as the mapping network for the Neuro-SM model of power transistors.…”
Abstract:In this paper, a novel neuro-space mapping (Neuro-SM) modeling approach for lager-signal transistors is proposed. A new structure of Neuro-SM model with capacitors and inductors is created to change the DC and AC characteristic of the model respectively. An additional current signal extracted with a novel nonlinear function is adopted to improve the large-signal characteristic of existing device models while remain the S-parameters unchanged. A step-by-step training method is developed for fast training of the proposed Neuro-SM model avoiding variables adjustment repeatedly. In addition, the modeling experiment for measurement data of LDMOS transistor demonstrate that the novel Neuro-SM method can accurately reflect the large-signal characteristics of transistor with simple operation process and enhance the accuracy of the existing model.
“…This technique uses neural networks to map the voltage or current signals of the existing device model into that of the device data. The Neuro-SM method can be applied to not only the simple DC and S-parameters modeling of nonlinear devices, but also the complex large-signal modeling [6,7,8]. In [9], a dynamic neural network is used as the mapping network for the Neuro-SM model of power transistors.…”
Abstract:In this paper, a novel neuro-space mapping (Neuro-SM) modeling approach for lager-signal transistors is proposed. A new structure of Neuro-SM model with capacitors and inductors is created to change the DC and AC characteristic of the model respectively. An additional current signal extracted with a novel nonlinear function is adopted to improve the large-signal characteristic of existing device models while remain the S-parameters unchanged. A step-by-step training method is developed for fast training of the proposed Neuro-SM model avoiding variables adjustment repeatedly. In addition, the modeling experiment for measurement data of LDMOS transistor demonstrate that the novel Neuro-SM method can accurately reflect the large-signal characteristics of transistor with simple operation process and enhance the accuracy of the existing model.
“…Moreover, an extra simulator must be involved in the process and linked to the optimization algorithm. Another problem is that performance of SM algorithm depends heavily on the selection of the SM transformations used to construct the surrogate [21].…”
A robust algorithm for simulation-driven design optimization of microwave structures evaluated using CST MICROWAVE STUDIO V R is described. The algorithm exploits gradient information obtained using adjoint sensitivity (if available) or finite differentiation. It also uses trust region approach that ensures good convergence properties and improves the overall performance. The efficiency of our approach is demonstrated using several examples of microwave structures. We also discuss an extension of the algorithm where the sensitivity of complex-valued responses rather than of the real-valued ones is used. A performance comparison with other optimization techniques is also provided.
“…A number of techniques for modeling and simulationdriven design of microwave structures have emerged over the recent years, including the methods that exploit artificial neural networks [8,[9][10][11][12], fuzzy systems [13], kriging [14], or multidimensional Cauchy approximation [15], as well as surrogate-based techniques such as space mapping (SM) [16][17][18][19][20][21][22][23]24], simulation-based tuning [25,26], and combination of both [27,28]. The last three approaches offer computationally efficient design optimization where, under certain circumstances, a satisfactory design can be obtained after a few high-fidelity (or fine) EM simulations of the structure of interest [16].…”
Simulation-based optimization has become an important design tool in microwave engineering. However, using electromagnetic (EM) solvers in the design process is a challenging task, primarily due to a high-computational cost of an accurate EM simulation. In this article, we present a review of EM-based design optimization techniques exploiting response-corrected physically based low-fidelity models. The surrogate models created through such a correction can be used to yield a reasonable approximation of the optimal design of the computationally expensive structure under consideration (high-fidelity model). Several approaches using this idea are reviewed including output space mapping, manifold mapping, adaptive response correction, and shape-preserving response prediction. A common feature of these methods is that they are easy to implement and computationally efficient. Application examples are provided.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.