Fast Multi-Objective Optimization of Multi-Parameter Antenna Structures Based on Improved BPNN Surrogate Model

Dong, Jian; Qin, Wenwen; Wang, Meng

doi:10.1109/access.2019.2920945

Cited by 76 publications

(56 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Undoubtedly, the most popular class of surrogates are data‐driven models. The widely used techniques include kriging, Gaussian process regression, artificial neural networks, or support vector regression …”

Section: Multiobjective Design Frameworkmentioning

confidence: 99%

“…54,55 Undoubtedly, the most popular class of surrogates are data-driven models. The widely used techniques include kriging, 56 Gaussian process regression, 57 artificial neural networks, 58 or support vector regression. 59 Some of the recently proposed surrogate-assisted MO frameworks 54,60,61 utilize the surrogate to yield the initial approximation of the Pareto set.…”

Section: Multiobjective Design Frameworkmentioning

confidence: 99%

See 1 more Smart Citation

Accelerated multiobjective design of miniaturized microwave components by means of nested kriging surrogates

Pietrenko‐Dabrowska

Koziel

2020

Int J RF Microw Comput Aided Eng

View full text Add to dashboard Cite

Design of microwave components is an inherently multiobjective task. Often, the objectives are at least partially conflicting and the designer has to work out a suitable compromise. In practice, generating the best possible trade‐off designs requires multiobjective optimization, which is a computationally demanding task. If the structure of interest is evaluated through full‐wave electromagnetic (EM) analysis, the employment of widely used population‐based metaheuristics algorithms may become prohibitive in computational terms. This is a common situation for miniaturized components, where considerable cross‐coupling effects make traditional representations (eg, network equivalents) grossly inaccurate. This article presents a framework for accelerated EM‐driven multiobjective design of compact microwave devices. It adopts a recently reported nested kriging methodology to identify the parameter space region containing the Pareto front and to render a fast surrogate, subsequently used to find the first approximation of the Pareto set. The final trade‐off designs are produced in a separate, surrogate‐assisted refinement process. Our approach is demonstrated using a three‐section impedance matching transformer designed for the best matching and the minimum footprint area. The Pareto set is generated at the cost of only a few hundred of high‐fidelity EM simulations of the transformer circuit despite a large number of geometry parameters involved.

show abstract

Section: Multiobjective Design Frameworkmentioning

confidence: 99%

Section: Multiobjective Design Frameworkmentioning

confidence: 99%

Accelerated multiobjective design of miniaturized microwave components by means of nested kriging surrogates

Pietrenko‐Dabrowska

Koziel

2020

Int J RF Microw Comput Aided Eng

View full text Add to dashboard Cite

show abstract

“…Fortunately, the recently developed surrogate-based optimization techniques [5][6][7][8][9][10][11][12] have proven to be more computationally efficient compared with conventional EM-driven simulations. Compared with traditional EM-driven approaches, surrogate-based optimization techniques construct a mathematical mapping between the antenna dimensions and antenna performance, thereby greatly reducing EM simulations and the computational cost.…”

Section: Introductionmentioning

confidence: 99%

“…Compared with traditional EM-driven approaches, surrogate-based optimization techniques construct a mathematical mapping between the antenna dimensions and antenna performance, thereby greatly reducing EM simulations and the computational cost. Different surrogate models are proposed for antenna optimizations, such as Kriging [5,6], Gaussian Process (GP) [7,8], and artificial neural networks (ANNs) [9][10][11][12]. The Kriging method used in References [5,6] is essentially an interpolation method with poor generalization and the prediction accuracy depends mostly on the initial sampling, which may cause the model to either stop prematurely or search too locally [13].…”

Section: Introductionmentioning

confidence: 99%

“…ANNs [9][10][11][12] are flexible mathematical structures which are capable of identifying complex nonlinear relationships between input and output data sets. Generally, backward propagation neural networks (BPNNs) [9,10] and radial basis function neural networks (RBFNNs) [11,12] are two kinds of typical models among ANNs. Compared to BPNNs, RBFNNs converge much faster as they have both a supervised and unsupervised component to their learning [12,15].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Fast Multi-Objective Antenna Optimization Based on RBF Neural Network Surrogate Model Optimized by Improved PSO Algorithm

Dong

Wang

2019

Applied Sciences

Self Cite

View full text Add to dashboard Cite

In this paper, a radial basis function neural network (RBFNN) surrogate model optimized by an improved particle swarm optimization (PSO) algorithm is developed to reduce the computation cost of traditional antenna design methods which rely on high-fidelity electromagnetic (EM) simulations. Considering parameters adjustment and update mechanism simultaneously, two modifications are proposed in this improved PSO. First, time-varying learning factors are designed to balance exploration and exploitation ability of particles in the search space. Second, the local best information is added to the updating process of particles except for personal and global best information for better population diversity. The improved PSO is applied to train RBFNN for determining optimal network parameters. As a result, the constructed improved PSO-RBFNN model can be used as a surrogate model for antenna performance prediction with better network generalization capability. By integrating the improved PSO-RBFNN surrogate model with multi-objective evolutionary algorithms (MOEAs), a fast multi-objective antenna optimization framework for multi-parameter antenna structures is then established. Finally, a Pareto-optimal planar miniaturized multiband antenna design is presented, demonstrating that the proposed model provides better prediction performance and considerable computational savings compared to those previously published approaches.

show abstract

Low‐cost multi‐criteria design optimization of compact microwave passives using constrained surrogates and dimensionality reduction

Koziel

Pietrenko‐Dabrowska

Al‐Hasan

2020

Int J Numerical Modelling

View full text Add to dashboard Cite

Design of contemporary microwave circuits is a challenging task. Typically, it has to take into account several performance requirements and constraints.The design objectives are often conflicting and their simultaneous improvement may not be possible; instead, compromise solutions are to be sought. Representative examples are miniaturized microwave passives where reduction of the circuit size has a detrimental effect on its electrical characteristics. Acquiring information about the best possible design trade-offs is invaluable for the designer, yet it entails computationally expensive multi-objective optimization (MO). MO is typically conducted using population-based metaheuristic algorithms, the cost of which might be extremely high. If the circuit performance is evaluated using full-wave electromagnetic (EM) analysis, this cost is often prohibitive. A workaround is the employment of fast surrogate models, and a number of surrogate-assisted frameworks have been proposed in the literature. Unfortunately, a construction of reliable surrogates is hindered in higher dimensional parameter spaces. The recently proposed constrained modeling mitigates this issue to a certain extent by restricting the modeling process to the region containing the Pareto front to be found. This work proposes a novel surrogate-based MO technique that involves constrained modeling and explicit reduction of the surrogate domain dimensionality. The latter is achieved through the spectral analysis of the extreme Pareto-optimal design set obtained by local search routines. Our methodology is validated using a 15-parameter impedance-matching transformer with the Pareto set identified at the cost of a few hundred EM analyses of the circuit. The numerical experiments also demonstrate a significant reduction of the optimization cost as compared to the state-of-the-art surrogate-assisted MO methods.

show abstract

Fast Multi-Objective Optimization of Multi-Parameter Antenna Structures Based on Improved BPNN Surrogate Model

Cited by 76 publications

References 28 publications

Accelerated multiobjective design of miniaturized microwave components by means of nested kriging surrogates

Accelerated multiobjective design of miniaturized microwave components by means of nested kriging surrogates

Fast Multi-Objective Antenna Optimization Based on RBF Neural Network Surrogate Model Optimized by Improved PSO Algorithm

Low‐cost multi‐criteria design optimization of compact microwave passives using constrained surrogates and dimensionality reduction

Contact Info

Product

Resources

About